[Bug] ESP32-S3-Box Lite only hears wakeword. Nothing else. 仅限唤醒词。其他什么都没有。
Author: kerrylandCreated Sep 16, 2026Updated Sep 17, 2026
Labelsbug
Before submitting
- I have tested the latest applicable firmware and searched for an existing issue reporting the same problem.
Device or development board
ESP32-S3-Box Lite
Firmware version
2.5.0
Firmware source or filename
https://github.com/78/xiaozhi-esp32/releases/download/v2.5.0/v2.5.0_espressif-esp-box-lite.zip
Problem and reproduction steps
My device recognises the wakeword, but never "hears" anything else I say. It happens every time. Logs are attached.
If I press the middle button, I also see this in the logs:
I (24607) StateMachine: State: connecting -> listening...but nothing is heard when I speak.
Obviously not a problem with the microphone, because the wakeword works.
Google translate, if it is helpful:
我的设备能识别唤醒词，但无法“听到”我说的其他任何内容。 每次都会出现这种情况。已附上日志。
如果我按下中间的按钮，日志中也会显示：
I (24607) StateMachine: State: connecting -> listening……但在我说话时，设备却没有任何反应。
显然问题不在麦克风上，因为唤醒词是有效的。
Logs or other evidence
I (417) LcdDisplay: Turning display on
I (2017) esp_netif_handlers: sta ip: 192.168.1.239, mask: 255.255.255.0, gw: 192.168.1.1
I (2017) WifiStation: Got IP: 192.168.1.239
I (2017) WifiBoard: Connected to WiFi: codftw
I (2017) Application: Network connected
I (2017) StateMachine: State: starting -> activating
I (2027) Assets: The storage free size is 21952 KB
I (2027) Assets: The assets map size is 320 KB (partition 8192 KB)
I (2057) Assets: The checksum calculation time is 16 ms
I (2057) MODEL_LOADER: Successfully load srmodels
I (2057) Assets: Refreshing display theme...
I (2057) Assets: Set hide_subtitle to false
I (2057) Ota: Current version: 2.5.0
I (2067) wifi:<ba-add>idx:1 (ifx:0, 04:42:1a:c4:ae:28), tid:0, ssn:1, winSize:64
I (3647) HttpClient: Established new connection to api.tenclass.net:443 protocol=https cost=1580
I (4327) HttpClient: HTTP connection closed
I (4337) Ota: Current is the latest version
I (4337) Ota: Running partition: ota_0
I (4337) MQTT: Connecting to endpoint api.tenclass.net:8883
I (4337) MQTT: Resolved MQTT host api.tenclass.net -> 47.76.65.170
I (6277) MQTT: Connected to endpoint
I (6277) Application: Activation done
I (6277) SystemInfo: free sram: 155939 minimal sram: 50740
I (6277) StateMachine: State: activating -> idle
I (6287) WifiStation: Setting WiFi power save level: LOW_POWER (MAX_MODEM)
I (6287) wifi:Set ps type: 2, coexist: 0
I (6297) AfeAudioEngine: Model 0: wn9_mycroft_tts
I (6297) WakeWordCache: Allocated 64000 bytes in PSRAM
I (6307) AFE_CONFIG: Set WakeNet Model: wn9_mycroft_tts
I (6307) AfeAudioEngine: Before AFE create: free=164275 min=50740 largest=98304
MC Quantized wakenet9: wakenet9l_tts1h8_Mycroft_3_0.625_0.629, tigger:v4, mode:4, p:0, (May 25 2026 15:51:00)
I (6477) AFE: AFE Version: (2MIC_V251128)
I (6477) AFE: Input PCM Config: total 3 channels(2 microphone, 1 playback), sample rate:16000
I (6487) AFE: AFE Pipeline: [input] -> |AEC(FD_LOW_COST, NLP_VERY_AGGRESSIVE)| -> |SE(BSS)| -> |VAD(WebRTC)| -> |WakeNet(wn9_mycroft_tts,)| -> [output]
I (6497) AFE: AFE Version: (2MIC_V251128)
I (6507) AFE: Input PCM Config: total 3 channels(2 microphone, 1 playback), sample rate:16000
I (6507) AFE: AFE Pipeline: [input] -> |SE(BSS)| -> |VAD(WebRTC)| -> [output]
I (6517) AfeAudioEngine: Initialized FD AFE, detector: WakeNet, NS: off, feed: 1024, fetch: 512
I (6527) AfeAudioEngine: After AFE create: free=145287 min=35828 largest=98304
I (6537) I2S_IF: Paired data: 0x3fceeb94, current mode: playback, paired in_enable: 0, paired out_enable: 0
I (6547) I2S_IF: STD: TX, data_bit: 16, slot_bit: 16, ws_width: 16, slot_mode: MONO, slot_mask: 0x1
I (6547) I2S_IF: STD: TX, sample_rate_hz: 16000, mclk_multiple: 256, clk_src: 6
I (6557) I2S_IF: Current mode(record) and peer mode have same sample_rate 16000, channel 2, bits_per_sample 16
I (6567) I2S_IF: TDM: RX, data_bit: 16, slot_bit: 16, ws_width: 16, total_slot: 2, slot_mask: 0x3
I (6577) I2S_IF: TDM: RX, sample_rate_hz: 16000, mclk_multiple: 256, clk_src: 6
I (6567) Adev_Codec: Open codec device OK
I (6587) AudioCodec: Set output enable to true
I (6597) Adev_Codec: Open codec device OK
I (6597) AudioCodec: Set input enable to true
I (13937) Application: Wake word detected: Mycroft (state: 3)
I (13947) StateMachine: State: idle -> connecting
I (13947) WifiStation: Setting WiFi power save level: PERFORMANCE (NONE)
I (13947) wifi:Set ps type: 0, coexist: 0
I (14187) AfeAudioEngine: Encoded wake word into 33 packets in 246 ms
I (14287) MQTT: Session ID: 0c6df374
I (14297) WifiStation: Setting WiFi power save level: PERFORMANCE (NONE)
I (14297) wifi:Set ps type: 0, coexist: 0
W (14297) Application: Server sample rate 24000 does not match device output sample rate 16000, resampling may cause distortion
I (14307) Application: Wake word detected: Mycroft
I (14347) StateMachine: State: connecting -> listening
I (14727) Application: >> Mycroft
I (14737) StateMachine: State: listening -> speaking
I (15027) Application: << Oh, Mycroft!
I (15207) AudioService: Resampling audio from 24000 to 16000
I (16667) Application: << Like the AI assistant from Sherlock Holmes stories?
I (19777) Application: << That's a cool name!
I (21397) Application: << What can I help you with today?
I (23387) StateMachine: State: speaking -> listening
I (32557) SystemInfo: free sram: 145139 minimal sram: 35828
I (42557) SystemInfo: free sram: 145411 minimal sram: 35828
I (52557) SystemInfo: free sram: 145391 minimal sram: 35828
I (62557) SystemInfo: free sram: 145411 minimal sram: 35828
I (72557) SystemInfo: free sram: 146055 minimal sram: 35828
I (82557) SystemInfo: free sram: 145139 minimal sram: 35828
I (90427) StateMachine: State: listening -> speaking
I (90737) Application: << Hey there, looks like it's been quiet for a bit.
I (93497) Application: << Take care, and catch you next time!
I (95597) StateMachine: State: speaking -> listening
I (95607) MQTT: Received goodbye message, session_id: 0c6df374
I (95617) MQTT: Closing audio channel, send_goodbye: 0
I (95617) WifiStation: Setting WiFi power save level: LOW_POWER (MAX_MODEM)
I (95617) wifi:Set ps type: 2, coexist: 0
I (95637) StateMachine: State: listening -> idle
I (105557) SystemInfo: free sram: 148987 minimal sram: 35828Optional photos, videos, or log files
No response
Source: 78/xiaozhi-esp32