feat(multinet): add API to modify speech commands online

2025-09-15 15:28:44 +08:00 · 2021-09-24 21:29:56 +08:00 · 2021-09-24 21:29:56 +08:00 · c3d120ac9d
commit c3d120ac9d
parent cfebc5f31a
8 changed files with 131 additions and 92 deletions
--- a/Kconfig.projbuild
+++ b/Kconfig.projbuild
@ -148,13 +148,13 @@ config EN_MULTINET3_SINGLE_RECOGNITION
    bool "english single recognition (MultiNet3)"
    depends on SR_ENGLISH && IDF_TARGET_ESP32

-config EN_MULTINET5_SINGLE_RECOGNITION
-    bool "english single recognition (MultiNet5)"
-    depends on SR_ENGLISH && IDF_TARGET_ESP32S3
-
 config EN_MULTINET5_SINGLE_RECOGNITION_QUANT8
    bool "english single recognition (MultiNet5 quantized with 8-bit)"
    depends on SR_ENGLISH && IDF_TARGET_ESP32S3
+    
+config EN_MULTINET5_SINGLE_RECOGNITION
+    bool "english single recognition (MultiNet5)"
+    depends on SR_ENGLISH && IDF_TARGET_ESP32S3

 config CN_MULTINET2_SINGLE_RECOGNITION
    bool "chinese single recognition (MultiNet2)"
@ -685,162 +685,162 @@ config CN_SPEECH_COMMAND_ID99
 config EN_SPEECH_COMMAND_ID0
    string "ID0"
    depends on SR_ENGLISH
-    default "T,f,L, ,M,m, ,c, ,q,b,K"
+    default "TfL Mm c qbK"

 config EN_SPEECH_COMMAND_ID1
    string "ID1"
    depends on SR_ENGLISH
-    default "S,g,l, ,c, ,S,e,l"
+    default "Sgl c Sel"

 config EN_SPEECH_COMMAND_ID2
    string "ID2"
    depends on SR_ENGLISH
-    default "P,L,d, ,N,o,Z, ,p,a,N,c,L"
+    default "PLd NoZ paNcL"

 config EN_SPEECH_COMMAND_ID3
    string "ID3"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,M,i, ,S,t,N,D,B,n,K,S"
+    default "TkN nN Mi StNDBnKS"

 config EN_SPEECH_COMMAND_ID4
    string "ID4"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,M,i, ,S,t,N,D,B,n,K,S"
+    default "TkN eF Mi StNDBnKS"

 config EN_SPEECH_COMMAND_ID5
    string "ID5"
    depends on SR_ENGLISH
-    default "h,i,c,S,T, ,V,n,L,Y,o,M"
+    default "hicST VnLYoM"

 config EN_SPEECH_COMMAND_ID6
    string "ID6"
    depends on SR_ENGLISH
-    default "L,b,c,S,T, ,V,n,L,Y,o,M"
+    default "LbcST VnLYoM"

 config EN_SPEECH_COMMAND_ID7
    string "ID7"
    depends on SR_ENGLISH
-    default "g,N,K,R,m,S, ,j,c, ,V,n,L,Y,o,M"
+    default "gNKRmS jc VnLYoM"

 config EN_SPEECH_COMMAND_ID8
    string "ID8"
    depends on SR_ENGLISH
-    default "D,g,K,R,m,S, ,j,c, ,V,n,L,Y,o,M"
+    default "DgKRmS jc VnLYoM"

 config EN_SPEECH_COMMAND_ID9
    string "ID9"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,j,c, ,T,m,V,m"
+    default "TkN nN jc TmVm"

 config EN_SPEECH_COMMAND_ID10
    string "ID10"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,j,c, ,T,m,V,m"
+    default "TkN eF jc TmVm"

 config EN_SPEECH_COMMAND_ID11
    string "ID11"
    depends on SR_ENGLISH
-    default "M,d,K, ,M,m, ,c, ,T,m"
+    default "MdK Mm c Tm"

 config EN_SPEECH_COMMAND_ID12
    string "ID12"
    depends on SR_ENGLISH
-    default "M,d,K, ,M,m, ,c, ,K,n,F,m"
+    default "MdK Mm c KnFm"

 config EN_SPEECH_COMMAND_ID13
    string "ID13"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,j,c, ,L,i,T"
+    default "TkN nN jc LiT"

 config EN_SPEECH_COMMAND_ID14
    string "ID14"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,j,c, ,L,i,T"
+    default "TkN eF jc LiT"

 config EN_SPEECH_COMMAND_ID15
    string "ID15"
    depends on SR_ENGLISH
-    default "p,d,N,q, ,j,c, ,K,c,L,k, ,T,o, ,R,f,D"
+    default "pdNq jc KcLk To RfD"

 config EN_SPEECH_COMMAND_ID16
    string "ID16"
    depends on SR_ENGLISH
-    default "p,d,N,q, ,j,c, ,K,c,L,k, ,T,o, ,G,R,m,N"
+    default "pdNq jc KcLk To GRmN"

 config EN_SPEECH_COMMAND_ID17
    string "ID17"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,e,L, ,j,c, ,L,i,T,S"
+    default "TkN nN eL jc LiTS"

 config EN_SPEECH_COMMAND_ID18
    string "ID18"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,e,L, ,j,c, ,L,i,T,S"
+    default "TkN eF eL jc LiTS"

 config EN_SPEECH_COMMAND_ID19
    string "ID19"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,j,c, ,f,R, ,K,c,N,D,g,s,c,N,k"
+    default "TkN nN jc fR KcNDgscNk"

 config EN_SPEECH_COMMAND_ID20
    string "ID20"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,j,c, ,f,R, ,K,c,N,D,g,s,c,N,k"
+    default "TkN eF jc fR KcNDgscNk"

 config EN_SPEECH_COMMAND_ID21
    string "ID21"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,S,g,K,S,T,m,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To SgKSTmN DgGRmZ"

 config EN_SPEECH_COMMAND_ID22
    string "ID22"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,S,f,V,c,N,T,m,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To SfVcNTmN DgGRmZ"

 config EN_SPEECH_COMMAND_ID23
    string "ID23"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,d,T,m,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To dTmN DgGRmZ"

 config EN_SPEECH_COMMAND_ID24
    string "ID24"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,N,i,N,T,m,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To NiNTmN DgGRmZ"

 config EN_SPEECH_COMMAND_ID25
    string "ID25"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm DgGRmZ"

 config EN_SPEECH_COMMAND_ID26
    string "ID26"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,W,c,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm WcN DgGRmZ"

 config EN_SPEECH_COMMAND_ID27
    string "ID27"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,T,o, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm To DgGRmZ"

 config EN_SPEECH_COMMAND_ID28
    string "ID28"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,v,R,m, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm vRm DgGRmZ"

 config EN_SPEECH_COMMAND_ID29
    string "ID29"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,F,e,R, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm FeR DgGRmZ"

 config EN_SPEECH_COMMAND_ID30
    string "ID30"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,F,i,V, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm FiV DgGRmZ"

 config EN_SPEECH_COMMAND_ID31
    string "ID31"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,S,g,K,S, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm SgKS DgGRmZ"

 config EN_SPEECH_COMMAND_ID32
    string "ID32"
--- a/acoustic_algorithm/libesp_audio_processor.a
+++ b/acoustic_algorithm/libesp_audio_processor.a
--- a/lib/libdl_lib_esp32s3.a
+++ b/lib/libdl_lib_esp32s3.a
--- a/lib/libmultinet.a
+++ b/lib/libmultinet.a
--- a/lib/libwakenet.a
+++ b/lib/libwakenet.a
--- a/main/Kconfig
+++ b/main/Kconfig
@ -148,13 +148,13 @@ config EN_MULTINET3_SINGLE_RECOGNITION
    bool "english single recognition (MultiNet3)"
    depends on SR_ENGLISH && IDF_TARGET_ESP32

-config EN_MULTINET5_SINGLE_RECOGNITION
-    bool "english single recognition (MultiNet5)"
-    depends on SR_ENGLISH && IDF_TARGET_ESP32S3
-
 config EN_MULTINET5_SINGLE_RECOGNITION_QUANT8
    bool "english single recognition (MultiNet5 quantized with 8-bit)"
    depends on SR_ENGLISH && IDF_TARGET_ESP32S3
+    
+config EN_MULTINET5_SINGLE_RECOGNITION
+    bool "english single recognition (MultiNet5)"
+    depends on SR_ENGLISH && IDF_TARGET_ESP32S3

 config CN_MULTINET2_SINGLE_RECOGNITION
    bool "chinese single recognition (MultiNet2)"
@ -685,162 +685,162 @@ config CN_SPEECH_COMMAND_ID99
 config EN_SPEECH_COMMAND_ID0
    string "ID0"
    depends on SR_ENGLISH
-    default "T,f,L, ,M,m, ,c, ,q,b,K"
+    default "TfL Mm c qbK"

 config EN_SPEECH_COMMAND_ID1
    string "ID1"
    depends on SR_ENGLISH
-    default "S,g,l, ,c, ,S,e,l"
+    default "Sgl c Sel"

 config EN_SPEECH_COMMAND_ID2
    string "ID2"
    depends on SR_ENGLISH
-    default "P,L,d, ,N,o,Z, ,p,a,N,c,L"
+    default "PLd NoZ paNcL"

 config EN_SPEECH_COMMAND_ID3
    string "ID3"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,M,i, ,S,t,N,D,B,n,K,S"
+    default "TkN nN Mi StNDBnKS"

 config EN_SPEECH_COMMAND_ID4
    string "ID4"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,M,i, ,S,t,N,D,B,n,K,S"
+    default "TkN eF Mi StNDBnKS"

 config EN_SPEECH_COMMAND_ID5
    string "ID5"
    depends on SR_ENGLISH
-    default "h,i,c,S,T, ,V,n,L,Y,o,M"
+    default "hicST VnLYoM"

 config EN_SPEECH_COMMAND_ID6
    string "ID6"
    depends on SR_ENGLISH
-    default "L,b,c,S,T, ,V,n,L,Y,o,M"
+    default "LbcST VnLYoM"

 config EN_SPEECH_COMMAND_ID7
    string "ID7"
    depends on SR_ENGLISH
-    default "g,N,K,R,m,S, ,j,c, ,V,n,L,Y,o,M"
+    default "gNKRmS jc VnLYoM"

 config EN_SPEECH_COMMAND_ID8
    string "ID8"
    depends on SR_ENGLISH
-    default "D,g,K,R,m,S, ,j,c, ,V,n,L,Y,o,M"
+    default "DgKRmS jc VnLYoM"

 config EN_SPEECH_COMMAND_ID9
    string "ID9"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,j,c, ,T,m,V,m"
+    default "TkN nN jc TmVm"

 config EN_SPEECH_COMMAND_ID10
    string "ID10"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,j,c, ,T,m,V,m"
+    default "TkN eF jc TmVm"

 config EN_SPEECH_COMMAND_ID11
    string "ID11"
    depends on SR_ENGLISH
-    default "M,d,K, ,M,m, ,c, ,T,m"
+    default "MdK Mm c Tm"

 config EN_SPEECH_COMMAND_ID12
    string "ID12"
    depends on SR_ENGLISH
-    default "M,d,K, ,M,m, ,c, ,K,n,F,m"
+    default "MdK Mm c KnFm"

 config EN_SPEECH_COMMAND_ID13
    string "ID13"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,j,c, ,L,i,T"
+    default "TkN nN jc LiT"

 config EN_SPEECH_COMMAND_ID14
    string "ID14"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,j,c, ,L,i,T"
+    default "TkN eF jc LiT"

 config EN_SPEECH_COMMAND_ID15
    string "ID15"
    depends on SR_ENGLISH
-    default "p,d,N,q, ,j,c, ,K,c,L,k, ,T,o, ,R,f,D"
+    default "pdNq jc KcLk To RfD"

 config EN_SPEECH_COMMAND_ID16
    string "ID16"
    depends on SR_ENGLISH
-    default "p,d,N,q, ,j,c, ,K,c,L,k, ,T,o, ,G,R,m,N"
+    default "pdNq jc KcLk To GRmN"

 config EN_SPEECH_COMMAND_ID17
    string "ID17"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,e,L, ,j,c, ,L,i,T,S"
+    default "TkN nN eL jc LiTS"

 config EN_SPEECH_COMMAND_ID18
    string "ID18"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,e,L, ,j,c, ,L,i,T,S"
+    default "TkN eF eL jc LiTS"

 config EN_SPEECH_COMMAND_ID19
    string "ID19"
    depends on SR_ENGLISH
-    default "T,k,N, ,n,N, ,j,c, ,f,R, ,K,c,N,D,g,s,c,N,k"
+    default "TkN nN jc fR KcNDgscNk"

 config EN_SPEECH_COMMAND_ID20
    string "ID20"
    depends on SR_ENGLISH
-    default "T,k,N, ,e,F, ,j,c, ,f,R, ,K,c,N,D,g,s,c,N,k"
+    default "TkN eF jc fR KcNDgscNk"

 config EN_SPEECH_COMMAND_ID21
    string "ID21"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,S,g,K,S,T,m,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To SgKSTmN DgGRmZ"

 config EN_SPEECH_COMMAND_ID22
    string "ID22"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,S,f,V,c,N,T,m,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To SfVcNTmN DgGRmZ"

 config EN_SPEECH_COMMAND_ID23
    string "ID23"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,d,T,m,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To dTmN DgGRmZ"

 config EN_SPEECH_COMMAND_ID24
    string "ID24"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,N,i,N,T,m,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To NiNTmN DgGRmZ"

 config EN_SPEECH_COMMAND_ID25
    string "ID25"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm DgGRmZ"

 config EN_SPEECH_COMMAND_ID26
    string "ID26"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,W,c,N, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm WcN DgGRmZ"

 config EN_SPEECH_COMMAND_ID27
    string "ID27"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,T,o, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm To DgGRmZ"

 config EN_SPEECH_COMMAND_ID28
    string "ID28"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,v,R,m, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm vRm DgGRmZ"

 config EN_SPEECH_COMMAND_ID29
    string "ID29"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,F,e,R, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm FeR DgGRmZ"

 config EN_SPEECH_COMMAND_ID30
    string "ID30"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,F,i,V, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm FiV DgGRmZ"

 config EN_SPEECH_COMMAND_ID31
    string "ID31"
    depends on SR_ENGLISH
-    default "S,f,T, ,j,c, ,T,f,M,P,R,c,p,k, ,T,o, ,T,W,f,N,T,m, ,S,g,K,S, ,D,g,G,R,m,Z"
+    default "SfT jc TfMPRcpk To TWfNTm SgKS DgGRmZ"

 config EN_SPEECH_COMMAND_ID32
    string "ID32"
--- a/speech_command_recognition/tool/README
+++ b/speech_command_recognition/tool/README
@ -1,18 +1,48 @@
-multinet_g2p.py 脚本可以把文本转为可以被multinet识别的因素代号．
-1. 使用之前需要先安装en_g2p
-```
-pip install en_g2p
-```
+**multinet_g2p.py** is used to convert English phrase into phonemes which can be recognized by multinet．

-2. 运行脚本，将文本呢转化为所需音素代号
-```
-python multinet_g2p.py -t "hello word"
----
-hello word  ->  ['h', 'c', 'L', 'b', ' ', 'W', 'k', 'D']
+
+
+####  1. Install g2p_en, please refer to https://pypi.org/project/g2p-en/ 

+```
+pip install g2p_en
 ```

-3. 将对应的短语的音素代号，添加到menuconfig即可
+#### 2. Run multinet_g2p.py
+
+```
+python multinet_g2p.py -t "hello world,hi ESP;turn on the light;turn off the light"
+
+------
+in: hello world,hi ESP;turn on the light;turn off the light
+out: hcLb WkLD,hi fST;TkN nN jc LiT;TkN eF jc LiT;
+```
+
+#### 3. Add speech commands 
+
+##### 3.1 add speech commands by menuconfig
+
+```
+idf.py menuconfig
+ESP Speech Recognition -> Add speech commands
+```



+##### 3.2 add speech commands by reset function
+
+```
+// Function definition
+// typedef void (*esp_mn_iface_op_reset_t)(model_iface_data_t *model_data, char *command_str, char *err_phrase_id);
+
+// "," is used to split different phrase with same command id
+// ";" is used to split different command id
+char *new_commands_str="hcLb WkLD,hi fST;TkN nN jc LiT;TkN eF jc LiT;"  //
+char err_id[256];
+multinet->reset(model_data, new_commands_str, err_id);
+// hello world,hi ESP -> commond id=0
+// turn on the light -> commond id=1
+// turn off the light -> commond id=2
+```
+
+
--- a/speech_command_recognition/tool/multinet_g2p.py
+++ b/speech_command_recognition/tool/multinet_g2p.py
@ -5,19 +5,28 @@ import pandas

 def english_g2p(text, alphabet=None):
    g2p = G2p()
-    labels = g2p(text)
-    out = []
+    out = ""

    if alphabet is None:
        alphabet={"AE1": "a", "N": "N", " ": " ", "OW1": "b", "V": "V", "AH0": "c", "L": "L", "F": "F", "EY1": "d", "S": "S", "B": "B", "R": "R", "AO1": "e", "D": "D", "AH1": "c", "EH1": "f", "OW0": "b", "IH0": "g", "G": "G", "HH": "h", "K": "K", "IH1": "g", "W": "W", "AY1": "i", "T": "T", "M": "M", "Z": "Z", "DH": "j", "ER0": "k", "P": "P", "NG": "l", "IY1": "m", "AA1": "n", "Y": "Y", "UW1": "o", "IY0": "m", "EH2": "f", "CH": "p", "AE0": "a", "JH": "q", "ZH": "r", "AA2": "n", "SH": "s", "AW1": "t", "OY1": "u", "AW2": "t", "IH2": "g", "AE2": "a", "EY2": "d", "ER1": "k", "TH": "v", "UH1": "w", "UW2": "o", "OW2": "b", "AY2": "i", "UW0": "o", "AH2": "c", "EH0": "f", "AW0": "t", "AO2": "e", "AO0": "e", "UH0": "w", "UH2": "w", "AA0": "n", "AY0": "i", "IY2": "m", "EY0": "d", "ER2": "k", "OY2": "u", "OY0": "u"}

-    for item in labels:
-        if item not in alphabet:
-            print("skip %s, not found in alphabet")
-            continue
-        else:
-            out.append(alphabet[item])
-    print(text, " -> ", out)
+    text_list = text.split(";")
+    for item in text_list:
+        item = item.split(",")
+        for phrase in item:
+            labels = g2p(phrase)
+            for char in labels:
+                if char not in alphabet:
+                    print("skip %s, not found in alphabet")
+                    continue
+                else:
+                    out += alphabet[char]
+            if phrase != item[-1]:
+                out += ','
+        out += ";"
+
+    print("in:", text)
+    print("out:", out)
    
    return out