ESP32-S3-BOX: How to get phoneme for custom voice command

[tushar@esp]

Hi, we started to experiment on ESP32-S3-BOX with inbuilt voice recognition.
We found below string with voice phoneme in the example from https://github.com/espressif/esp-box.

Code: Select all

[b]// cmd,                       Language,      id, string,                   phoneme,           next ref[/b]
    {SR_CMD_LIGHT_ON, SR_LANG_EN, 0, "Turn On the Light", "TkN nN jc LiT", {NULL}},
    {SR_CMD_LIGHT_ON, SR_LANG_EN, 0, "Switch On the Light", "SWgp nN jc LiT", {NULL}},
    {SR_CMD_LIGHT_OFF, SR_LANG_EN, 0, "Switch Off the Light", "SWgp eF jc LiT", {NULL}},
    {SR_CMD_LIGHT_OFF, SR_LANG_EN, 0, "Turn Off the Light", "TkN eF jc LiT", {NULL}},
.
.
.

when we create custom voice command, mobile application is creating phoneme and sending to device as per my understating from the example code.
If we want to implement the direct custom command with example, can someone please help to know how we can generate phoneme for desired string?
I tried to search and tried using some online convertor but the device is not recognizing that string as voice. Is there any specific phoneme conversation is required to make custom voice command?

[ESP_Rachel]

We have a tool that can convert word phrases into phoneme that speech recognition model can recognized. The tool has been deployed in the mobile App server.
Please refer to: https://github.com/espressif/esp-sr/tre ... ition/tool

[tushar@esp]

Okay - Got it.
thanks for the quick response.
IMO, It should be mentioned regarding this tool in https://github.com/espressif/esp-sr/README.md.