I try client-sdk-esp32, the sample project minimal...
# sdk-native
c
I try client-sdk-esp32, the sample project minimal, Here is the sample log: I (11542) AUD_PIPE_NEGO: Path mask 1 select sink:0 format 1398100047 I (11548) AUD_SRC: Get rate:16000, ch:1, bits:16 I (11553) ESP_GMF_TASK: Waiting to run... [tsk:aenc_0-0x3fcd33e0, wk:0x0, run:0] I (11559) I2S_IF: channel mode 2 bits:16/16 channel:4 mask:3 I (11559) livekit_example: Room state changed: Connected I (11565) I2S_IF: TDM Mode 0 bits:16/16 channel:4 sample_rate:16000 mask:3 I (11576) I2S_IF: channel mode 0 bits:16/16 channel:4 mask:3 I (11582) I2S_IF: STD Mode 1 bits:32/32 channel:4 sample_rate:16000 mask:0 I (11589) ES7210: Bits 16 I (11597) ES7210: Enable ES7210_INPUT_MIC1 I (11600) ES7210: Enable ES7210_INPUT_MIC2 I (11603) ES7210: Enable ES7210_INPUT_MIC3 I (11606) ES7210: Enable ES7210_INPUT_MIC4 I (11609) ES7210: Enable TDM mode I (11615) ES7210: Unmuted I (11615) Adev_Codec: Open codec device OK E (11615) MODEL_LOADER: Can not find model in partition table W (11620) AUD_AEC_SRC: No model to load W (11624) AFE_CONFIG: wakenet model not found. please load wakenet model... I (11674) AFE: AFE Version: (1MIC_V250121) I (11674) AFE: Input PCM Config: total 2 channels(1 microphone, 1 playback), sample rate:16000 I (11675) AFE: AFE Pipeline: [input] -> |AEC(SR_LOW_COST)| -> |VAD(WebRTC)| -> [output] I (11684) AUD_SRC: Start to fetch audio src data now I (11684) ESP_GMF_TASK: One times job is complete, del[wk:0x3c29bcbc, ctx:0x3c29ba48, label:aud_src_open] I (11697) ESP_GMF_PORT: ACQ IN, new self payload:0x3c29bcbc, port:0x3c2a0c00, el:0x3c29ba48-aud_src I (11706) ESP_GMF_AENC: Open, type:OPUS, acquire in frame: 640, out frame: 100 I (11713) ESP_GMF_TASK: One times job is complete, del[wk:0x3c314578, ctx:0x3c29bb24, label:aud_enc_open] The device connected to the room, but no voice output, and even I said somthing there is no related log changed in the console. I use default S3_Korvo_V2 board. The board is ok for mic and speaker My agent is working correct if I use the sandbox app. The embedded project is also setup with the sandbox configuration. Is there any suggestion I could find my embedded problem?
r
First off can you please thread your messages. It is hard to know which to respond to when there are multiple in the main channel for the same conversation. @gentle-refrigerator-18414 may have more good advise on the ESP issue you are having.
S3_Korvo_V2
is a good board to use. I've used our SDK on that specific board and have had no issues. Often agent is not dispatched because a room is getting reused. The agent will only get dispatched to a new room. Another common issue for folks just starting out is a mismatch between agent name in the agent and the dispatch.
g
Are you connecting using a token or sandbox id?
c
I use sandbox id. I use the same sandbox id in sandbox APP. After a success connection, from the session log, I could see agent dispatch, and make init. But if using ESP32 device, I could no see such agent dispatch . I thing the first problem is still about the connection. From the console log, device reported the room was connected. so I do not know how to debug current situation.
As I understand: 1. after conneciton, the agent should be dispatched. Is it true? One appearance I am not clear: from the session log, the participants is alway 1, even the the sandbox app could communicate with the agent. I think the two participants in the room could communicate with each other, whether the agent is not included for the participant?
Is it possible provide a minimal demo console log? My log is just waiting for I pasted status: audio is ready for use, but before the log, there is some warning for model is not existing in the partition, I think it is not important.