/*asrproxy启动不了,一般都是JSON格式不对,可以用 http://tools.jb51.net/code/jsonformat 这个工具检查JOSN是否存在错误。*/ { "key":"asrproxy.license", //授权文件路径
"log":{ "console_level":0, //输出控制台日志等级0-5(0:DEBUG, 1:INFO, 2:NOTICE, 3:WARNING, 4:CRIT, 5:CONSOLE) "file_level":0, //输出到文件日志等级0-5(0:DEBUG, 1:INFO, 2:NOTICE, 3:WARNING, 4:CRIT, 5:CONSOLE) "file_maxsize":100, //文件大于多少M就自动创建新的日志文件。 "file_number":10 //最大保留日志文件个数 },
"asr":{
"listen_ip":"127.0.0.1", "listen_port":9988, "storage":"record", //asr录音目录,调用asr时,设置了asr_params.recordfilename才会录音。 "acl":"*", //哪些IP可以访问,配置*任意IP都可以访问,多个IP用逗号隔开,如果不配置acl,通过127.0.0.1不需要配置在ACL里面也可以访问。
"interface": {
/*FunASR*/ /* "paraformer_realtime_16K": { "type": "funasr", "engine": "2pass", "quantize": true, "enable_itn": true, "itn-dir": "/ddt/asrproxy/funasr/fst_itn_zh", "vad-dir": "/ddt/asrproxy/funasr/speech_fsmn_vad_zh-cn-16k-common-onnx", "vad-quant": true, "punc-dir": "/ddt/asrproxy/funasr/punc_ct-transformer_zh-cn-common-vad_realtime-vocab272727-onnx", "model-dir": "/ddt/asrproxy/funasr/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-onnx", "online-model-dir": "/ddt/asrproxy/funasr/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online-onnx", "lm-dir": "", "hotword": "", "chunk-size": 800 }, */
/*阿里云私有云asr配置例子*/ /* "aliyun_test": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "type": "aliyun", "engine": "transcriber", //transcriber:流接口,recognizer:一句话接口 "appid": "default", "key": "default", "token": "default", "url": "ws://ip:8101/ws/v1", "secret": "default", "enable_intermediate_result": true, "enable_punctuation_prediction": false, "enable_inverse_text_normalization": false, "enable_semantic_sentence_detection": false }, */
/*阿里云实时语音识别配置例子 ,参数说明看https://help.aliyun.com/zh/isi/developer-reference/api-reference */ /* "aliyun_01": { "count": 0, //最大多少并发 "fault_threshold": 0, //连续错误多少次停用 "fault_try_interval": 600, //停用多久后,从新尝试用这个KEY "type": "aliyun", //识别引擎,通过配置指定调用的ASR类型 "engine": "transcriber", //transcriber:流接口,recognizer:一句话接口 "appid": "", //项目ID,详细看开通说明 https://help.aliyun.com/document_detail/71936.html "key": "", //AccessKey ID 详细看 https://help.aliyun.com/document_detail/324194.html "secret": "", //AccessKey Secret 详细看 https://help.aliyun.com/document_detail/69835.htm?spm=a2c4g.11186623.0.0.2f773f87BMksCK#section-iqi-ojp-tqk "enable_intermediate_result": true, //设置是否返回中间识别结果 "enable_punctuation_prediction": false, //设置是否在后处理中添加标点 "enable_inverse_text_normalization": false, //设置是否在后处理中执行数字转换 "enable_semantic_sentence_detection": false, //设置是否使用语义断句。 "customization_id": "", //设置定制模型 "vocabulary_id": "", //设置泛热词 "enable_ignore_sentencetimeout": true, //是否忽略实时识别中的单句识别超时, 默认是False "disfluency": true, //是否对识别文本进行顺滑(去除语气词,重复说等), 默认是False "speech_noise_threshold": 0.1, //噪音参数阈值,参数范围:[-1,1]取值越趋于-1:噪音被判定为语音的概率越大,取值越趋于+1:语音被判定为噪音的概率越大,该参数属高级参数, 调整需慎重并重点测试 "enable_sdk_log": false }, */
/*阿里云一句话识别配置例子 ,参数说明看 https://help.aliyun.com/zh/isi/developer-reference/restful-api-2 */ /* "aliyun_02": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "type": "aliyun", "engine": "recognizer", //transcriber:流接口,recognizer:一句话接口 "appid": "", "key": "", "secret": "", "enable_intermediate_result": true, "enable_punctuation_prediction": false, "enable_inverse_text_normalization": false, "enable_voice_detection": true //设置是否启动自定义静音检测 }, */
/*阿里云国际站 开通说明 https://www.alibabacloud.com/help/zh/isi/getting-started/start-here 接口文档看 https://www.alibabacloud.com/help/zh/isi/developer-reference/api-reference */ /* "aliyun_international": { "count": 0, //最大多少并发 "fault_threshold": 0, //连续错误多少次停用 "fault_try_interval": 600, //停用多久后,从新尝试用这个KEY "type": "aliyun", //识别引擎,通过配置指定调用的ASR类型 "engine": "transcriber", //transcriber:流接口,recognizer:一句话接口 "appid": "", //项目APPKEY,详细看开通说明 https://www.alibabacloud.com/help/zh/isi/getting-started/manage-projects "key": "", //AccessKey ID 看secret的说明 "secret": "", //AccessKey Secret 需要给予智能语音交换的权限 详细看 https://www.alibabacloud.com/help/zh/isi/activate-intelligent-speech-interaction-1 "url":"wss://nls-gateway-ap-southeast-1.aliyuncs.com/ws/v1", // 接口地址详细看 https://www.alibabacloud.com/help/zh/isi/developer-reference/api-reference "tokenurl":"http://nlsmeta.ap-southeast-1.aliyuncs.com", // tokenurl、regionid、version 这3个参数 国际站才需要配置 具体看 https://www.alibabacloud.com/help/zh/isi/getting-started/use-http-or-https-to-obtain-an-access-token "regionid":"ap-southeast-1", "version":"2019-02-28", "enable_intermediate_result": true, //设置是否返回中间识别结果 "enable_punctuation_prediction": false, //设置是否在后处理中添加标点 "enable_inverse_text_normalization": false, //设置是否在后处理中执行数字转换 "enable_semantic_sentence_detection": false, //设置是否使用语义断句。 "customization_id": "", //设置定制模型 "vocabulary_id": "", //设置泛热词 "enable_ignore_sentencetimeout": true, //是否忽略实时识别中的单句识别超时, 默认是False "disfluency": true, //是否对识别文本进行顺滑(去除语气词,重复说等), 默认是False "speech_noise_threshold": 0.1, //噪音参数阈值,参数范围:[-1,1]取值越趋于-1:噪音被判定为语音的概率越大,取值越趋于+1:语音被判定为噪音的概率越大,该参数属高级参数, 调整需慎重并重点测试 "enable_sdk_log": false }, */
//讯飞语音听写接口,开通地址 https://www.xfyun.cn/services/voicedictation ,如果需要启用抢话功能,需要开通 "动态修正" /* "xfyun_0": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "connect_timeout":3000, //建立ws连接最大等待时间 "accent" : "mandarin", //方言,这几个参数的说明看接口文档 https://www.xfyun.cn/doc/asr/voicedictation/API.htm "domain" : "iat", //应用领域 "language" : "zh_cn", //语种 "dwa" : "wpgs", //动态修正 "pd" : "", //领域个性化参数 "rlang" : "zh-cn", //繁体简体 "ptt" : 1, //是否开启标点符号添加 "nunum" : 1, //将返回结果的数字格式规则为阿拉伯数字格式 "type": "xfyun", //xfyun 标识这个配置试用讯飞云ASR引擎 "appid": "", //APPID "key": "", //APIKey "secret": "" //APISecret }, */
/* "xfyun_1": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "type": "xfyun", "url": "ws://ip:1078/iat", //私有云地址,公有云默认地址ws://iat-api.xfyun.cn/v2/iat,可以不配置。 "appid": "", "key": "", "secret": "" }, */
/* "mrcp": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "type": "mrcp", "engine": "uni2", "grammar": "", "appid": "", "key": "", "secret": "" }, */
//实时语音识别(websocket)https://cloud.tencent.com/document/product/1093/48982 /* "tencent": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "type": "tencent", "engine": "8k_zh", //对应文档 engine_model_type 参数说明 "needvad":"1", "filter_dirty":"0", "filter_modal":"0", "filter_punc":"0", "convert_num_mode":"1", "noise_threshold":"0", "hotword_id":"", "customization_id":"", "emotion_recognition":"0", "appid": "", "key": "", "secret": "" }, */ //获取取帐号名、帐号ID、用户名、用户ID、项目名称、项目ID 方法看这里 https://support.huaweicloud.com/api-iam/iam_17_0002.html //区域 列表 https://developer.huaweicloud.com/endpoint?IAM //可用终端节点 https://support.huaweicloud.com/api-sis/sis_03_0004.html 写这个说明的时以下节点可用 //华北-北京一 cn-north-1, 华北-北京四 cn-north-4, 华东-上海一 cn-east-3 //账户信息 请看 https://support.huaweicloud.com/api-sis/sis_03_0058.html Token认证的描述 /* "huawei": { "type": "huawei", "projectname": "", //项目名字实际含义是项目节点 我的凭证处获取,如果没有项目先创建项目 "projectid": "", //项目ID 我的凭证处获取 "add_punc": true, //表示是否在识别结果中添加标点 "digit_norm": true, //表示是否将语音中的数字识别为阿拉伯数字 "interim_results": true, //是否输出中间结果 "vocabulary_id": "", //热词表id 创建方法看 https://support.huaweicloud.com/api-sis/sis_03_0071.html "engine": "short-audio", //short-audio:一句话接口 https://support.huaweicloud.com/api-sis/sis_03_0095.html "key": "", //Access Key Id 我的凭证(访问密钥)处创建 "secret": "" //Secret Access Key 创建访问密钥时下载获取 }, */
/* 多方ASR */ /* "duofang": { "type": "duofang", "appid": "", "key": "", "secret": "", "im":true, //中间结果是否返回 "itn":true, //是否开启文本规整 "pct":true //是否添加标点 }, */
//思必驰语音识别 接口文档 https://help.talkinggenie.com/#/websocket_asr_vad /* "aispeech": { "type": "talkinggenie", "appid": "", //PID 产品ID "key": "", //PublicKey 说明文档 https://help.talkinggenie.com/#/api "secret": "" //SecretKey }, */
//百炼Paraformer https://help.aliyun.com/zh/model-studio/developer-reference/websocket-for-paraformer-real-time-service /* "dashscope": { "engine":"paraformer-realtime-8k-v2", //模型名称:"paraformer-realtime-v2",支持的模型请参考API详情 "vocabulary_id":"",//热词ID "disfluency_removal_enabled":false, //过滤语气词,默认关闭。 "language_hints":"", //指定识别语音中语言的代码列表。zh,yue,en,ja,ko 如果不填写则模型会自动识别语种。 "semantic_punctuation_enabled":false, //是否开启语义断句,服务默认开启语义断句,如果关闭则使用VAD (Voice Activity Detection)断句。 "inverse_text_normalization_enabled":true, //是否开启ITN(逆文本,Inverse Text Normalization),开启后(设置为True)中文数字会转换为阿拉伯数字。 "type": "dashscope", "key":"", //获取APIKEY 请看 https://help.aliyun.com/zh/model-studio/developer-reference/get-api-key }, */
//AppID/Token 请看https://www.volcengine.com/docs/6561/196768#q1%EF%BC%9A%E5%93%AA%E9%87%8C%E5%8F%AF%E4%BB%A5%E8%8E%B7%E5%8F%96%E5%88%B0%E4%BB%A5%E4%B8%8B%E5%8F%82%E6%95%B0appid%EF%BC%8Ccluster%EF%BC%8Ctoken%EF%BC%8Cauthorization-type%EF%BC%8Csecret-key-%EF%BC%9F //火山引擎ASR https://www.volcengine.com/docs/6561/196768 //一句话识别和流识别主要是计费方式不一样,一句话最大说话时间是60秒,流识别和大模型识别没限制。 "Bytedance": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "engine":"largemodel", //大模型:largemodel 一句话:volcengine_input_common 流识别:volcengine_streaming_common 非普通话识别: 设置为火山引擎后台的“Cluster ID” "enable_itn":true, //启用itn。文本规范化 (ITN) 是自动语音识别 (ASR) 后处理管道的一部分。 ITN 的任务是将 ASR 模型的原始语音输出转换为书面形式,以提高文本的可读性。 "enable_punc":true, //启用标点 "enable_ddc":false, //启用顺滑。*语义顺滑**是一种技术,旨在提高自动语音识别(ASR)结果的文本可读性和流畅性。这项技术通过删除或修改ASR结果中的不流畅部分,如停顿词、语气词、语义重复词等,使得文本更加易于阅读和理解。 "boosting_table_name":"", //热词词表名称 "boosting_table_id":"", //热词词表id,大模型识别才有这个配置 "type": "bytedance", "appid": "", "token": "" },
/* "asr配置名字:新加ASR配置会自动生效(1分钟之内)": { "说明": "已经存在的修改type,engine,appid,key,secret 不会生效,如果要修改这几个参数只能新添加一个ASR配置或者重启asrproxy程序,其他参数修改后自动生效" } */ }, "groups":{ //默认使用的ASR "default":{ "mode":0, //0:顺序使用,当使用数量等于count的时候切换下一个 1:循环使用 "enable": //启用的那些ASR配置 [ "改成要使用的ASR配置" ] },
//自定义组,可以话术中指定使用那组ASR配置。 "custom":{ "mode":0, "enable":[ ] } } }, "tts":{ "disable_caching":false,//是否禁用缓存 "storage": "tts", //tts缓存目录 "append_silence": 20, //tts后的声音文件前后追加静音 "trim_threshold_volume": 50, //tts后的声音文件前后小于多少音量认为是静音去除 "listen_ip":"127.0.0.1", "listen_port": 9989, "bgtts_thread_count":10,
"interface":{
/* "aliyun":{ "count":0, "fault_threshold":0, "fault_try_interval":600, "type":"aliyun", "appid":"", "key":"", "secret":"" }, */
/*阿里云国际站 开通说明 https://www.alibabacloud.com/help/zh/isi/getting-started/start-here 接口文档 https://www.alibabacloud.com/help/zh/isi/developer-reference/overview-of-speech-synthesis */ /* "aliyun_international": { "count": 0, //最大多少并发 "fault_threshold": 0, //连续错误多少次停用 "fault_try_interval": 600, //停用多久后,从新尝试用这个KEY "type": "aliyun", //TTS引擎,通过配置指定调用的TTS类型 "appid": "", //项目APPKEY,详细看开通说明 https://www.alibabacloud.com/help/zh/isi/getting-started/manage-projects "key": "", //AccessKey ID 看secret的说明 "secret": "", //AccessKey Secret 需要给予智能语音交换的权限 详细看 https://www.alibabacloud.com/help/zh/isi/activate-intelligent-speech-interaction-1 "url":"http://nls-gateway-ap-southeast-1.aliyuncs.com/stream/v1/tts", // 接口地址详细看 https://www.alibabacloud.com/help/zh/isi/developer-reference/restful-api-3 "tokenurl":"http://nlsmeta.ap-southeast-1.aliyuncs.com", // tokenurl、regionid、version 这3个参数 国际站才需要配置 具体看 https://www.alibabacloud.com/help/zh/isi/getting-started/use-http-or-https-to-obtain-an-access-token "regionid":"ap-southeast-1", "version":"2019-02-28" }, */
//阿里云智能语音平台的 cosyvoice(注意百练平台也有cosyvoice,请看下面dashscope配置说明) 接口文档 https://help.aliyun.com/zh/isi/developer-reference/speech-synthesis-1 /* "aliyun_cosyvoice": { "count": 0, "url":"ws://nls-gateway-cn-beijing.aliyuncs.com/ws/v1", //如果用cosyvoice的发音人,URL需要设置为这个 "fault_threshold": 0, "fault_try_interval": 600, "type": "aliyun", "appid": "", "key": "", "secret": "" }, */
/* "baidu":{ "connect_timeout":3000, "response_timeout":60000, "count":0, "fault_threshold":0, "fault_try_interval":600, "type":"baidu", "appid":"", "key":"", "secret":"" }, */
//讯飞语音合成接口,开通地址 https://console.xfyun.cn/services/tts /* "xfyun": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "rdn": "0", //合成音频数字发音方式,参数说明 https://www.xfyun.cn/doc/tts/online_tts/API.html "reg": "0", //设置英文发音方式: "connect_timeout": 3000, //建立ws连接最大等待时间 "response_timeout": 60000, //等待tts返回的最大时间 "type": "xfyun", //xfyun 标识这个配置试用讯飞云ASR引擎 "appid": "", //APPID "key": "", //APIKey "secret": "" //APISecret }, */
//华为语音合成 https://support.huaweicloud.com/api-sis/sis_03_0111.html /* "huawei": { "connect_timeout": 3000, "response_timeout": 60000, "type": "huawei", "projectname": "", //项目名字实际含义是项目节点 我的凭证处获取,如果没有项目先创建项目 "projectid": "", //项目ID 我的凭证处获取 "key": "", //Access Key Id 我的凭证(访问密钥)处创建 "secret": "" //Secret Access Key 创建访问密钥时下载获取 }, */
//思必驰语音识别 接口文档 https://help.talkinggenie.com/#/websocket_asr_vad /* "aispeech": { "type": "talkinggenie", "appid": "", //PID 产品ID "key": "", //PublicKey 说明文档 https://help.talkinggenie.com/#/api "secret": "" //SecretKey }, */
//百炼语音合成 只有cosyvoice支持声音复刻 介绍文档请看 https://help.aliyun.com/zh/model-studio/text-to-speech 和 https://help.aliyun.com/zh/model-studio/qwen-tts /* "dashscope": { "engine":"sambert", //sambert(https://help.aliyun.com/zh/model-studio/developer-reference/sambert-websocket-api) 或者 cosyvoice-v1(https://help.aliyun.com/zh/model-studio/cosyvoice-websocket-api) cosyvoice-v2(https://help.aliyun.com/zh/model-studio/cosyvoice-websocket-api#5fd1dd35c0qot) qwen-tts(https://help.aliyun.com/zh/model-studio/qwen-tts-api) "type": "dashscope", "key":"", //获取APIKEY 请看 https://help.aliyun.com/zh/model-studio/developer-reference/get-api-key }, */
//AppID/Token 请看https://www.volcengine.com/docs/6561/196768#q1%EF%BC%9A%E5%93%AA%E9%87%8C%E5%8F%AF%E4%BB%A5%E8%8E%B7%E5%8F%96%E5%88%B0%E4%BB%A5%E4%B8%8B%E5%8F%82%E6%95%B0appid%EF%BC%8Ccluster%EF%BC%8Ctoken%EF%BC%8Cauthorization-type%EF%BC%8Csecret-key-%EF%BC%9F //普通语音识别发音人https://www.volcengine.com/docs/6561/97465 //大模型语音识别发音人https://www.volcengine.com/docs/6561/1257544 //如果需要指定情感和语种方法,把发音人设置为 发音人.情感.语种 比如 BV700_streaming.customer_service.en //支持以下参数explicit_language、context_language、disable_markdown_filter、enable_latex_tn、split_sentence、pure_english_opt,参数的类型和用法请看火山官方文档 https://www.volcengine.com/docs/6561/79823 /* "Bytedance": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "engine":"largemodel", //大模型语音合成设置为largemodel,普通合成设置为空。 "type": "bytedance", "appid": "", "token": ""
} */
/* "mrcp": { "count": 0, "fault_threshold": 0, "fault_try_interval": 600, "type": "mrcp", "engine": "uni2", "appid": "", "key": "", "secret": "" }, */
//飞桨TTS模型 /* "paddlespeech": { "type": "paddlespeech", "front_conf":"paddlespeech_tts/front.conf", "am_model":"paddlespeech_tts/weights/fastspeech2_csmsc.onnx", "voc_model":"paddlespeech_tts/weights/mb_melgan_csmsc.onnx", "thread_num":1 } */
}, "groups":{ "default":{ "mode":1, "enable":[ "改成要使用的TTS配置名" ] },
//自定义组,可以话术中指定使用那组TTS配置。 "custom":{ "mode":0, "enable":[ ] } } } }
|