devttl/xiaozhi-esp32
1
0
Fork 0
forked from xiaozhi/xiaozhi-esp32
Code Pull Requests Activity
733 Commits 1 Branch 67 Tags
7f332f120c0aaee7f5f63e6bef189ccd4fb8e217
Commit Graph

110 Commits

Author SHA1 Message Date
virgil
a601a5cbc1 feat: sensecap watcher add inference (#1312) 
* feat: Wake up when a person is detected

* fix: Solve the problem of no sound when using WakeWordInvoke

* fix: Solve the problem of triggering dialogue when the person has not left

* feat(vision): 优化视觉检测逻辑并增加配置接口

本次提交旨在优化视觉检测功能,使其行为更自然、更智能,并为用户提供灵活的配置选项。

主要更新包括:

1. 引入了更精细的检测状态机:

   - IDLE: 空闲状态,等待检测目标。

   - VALIDATING: 验证状态,在检测到目标后,持续一段时间(可配置)以确认其存在,防止误触发。

   - COOLDOWN: 冷却状态,在一次成功交互后进入,避免过于频繁的打扰。

2. 新增了用于配置视觉检测的 MCP 工具:

   - self.vision.get_detection_config: 获取当前的检测参数(阈值、冷却间隔、验证时长、目标类型)。

   - self.vision.set_detection_config: 允许用户动态修改这些参数,以适应不同场景。

3. 性能优化:

   - 增加了配置参数的内存缓存,避免了在检测循环中对 NVS 的频繁访问。

* feat: Inference using Model 4

* feat: default inference disable

* feat: version cmd change to output json

* fix: fix image display

* Fix include directives for esp_check and esp_app_desc

---------

Co-authored-by: Spencer <love4yzp@gmail.com>
Co-authored-by: Xiaoxia <terrence@tenclass.com>
 2025-10-20 21:18:44 +08:00
Xiaoxia
d3e7fee828 fix multiple wakenet words and custom wake word (#1226) 
* fix multiple wakenet words and custom wake word

* fix idf_component.yml
 2025-09-22 10:49:08 +08:00
Xiaoxia
5018f6c03a feat: build default assets instead of downloading and v2 tables for esp-hi, echoear (#1203) 
* fix: call flush ready on io ready

* eachear: update to v2 partition table but disable class Assets currently

* stop gif if previewing an image

* feat: build default assets instead of downloading

* version updates

* fix None error

* delay 1s before enter wifi config mode

* fix compiling with v1 partition table

---------

Co-authored-by: Xiaoxia <terrence.huang@tenclass.com>
 2025-09-16 15:24:24 +08:00
Xiaoxia
b413e3ec03 regenerate jpeg encoder (#1198) 
* regenerate jpeg encoder

* add README to gif/jpeg

* 开机启动显示开发板信息，提前启动event loop
 2025-09-16 05:00:02 +08:00
Xiaoxia
147d71b9f1 feat: add snapshot mcp tool (#1196) 
* use main task to execute tool calls

* feat: add snapshot mcp tool

* fix compiling errors

* 取消 audio input 的 pin core，core1留给显示，可能会对aec性能有影响

* update ml307 version

* remove v1 theme colors
 2025-09-14 15:16:49 +08:00
Terrence
f79506d58b feat: add manual upgrade firmware 2025-09-12 08:45:29 +08:00
Xiaoxia
83f6f8c703 Switch to 2.0 branch (#1152) 
* Adapt boards to v2 partition tables

* fix esp log error

* fix display style

* reset emotion after download assets

* fix compiling

* update assets default url

* Add user only tools

* Add image cache

* smaller cache and buffer, more heap

* use MAIN_EVENT_CLOCK_TICK to avoid audio glitches

* bump to 2.0.0

* fix compiling errors

---------

Co-authored-by: Xiaoxia <terrence.huang@tenclass.com>
 2025-09-04 15:41:28 +08:00
Xiaoxia
5d3f597137 Bump to v1.9.0 (#1157) 
* update v2 partition table readme

* feat: Add user only tool

* Add image cache

* smaller cache and buffer, more heap

* use MAIN_EVENT_CLOCK_TICK to avoid audio glitches

* fix: esp_psram_get_size not found in c3

* Bump to 1.9.0
 2025-09-04 12:30:26 +08:00
Xiaoxia
d09537ed5c Add V2 parition tables (#1137) 
* v1.9.0: update font icons, add mqtt reconnect

* Add v2 parition tables
 2025-08-29 09:04:23 +08:00
Xiaoxia
9ae34d3b45 feat: Add locales with OGG sounds (#1057) 
* fix jiuchuan build problem

* feat: Add locales with OGG sounds

* fix building echoear

* Support ogg files frame duration <= 60
 2025-08-12 18:41:00 +08:00
Xiaoxia
363073658a fix: OTA buffer overflow caused by slow SetChatMessage (#1031) 2025-08-06 01:16:52 +08:00
Terrence
d58d55cfe7 Bump to 1.8.5 2025-08-02 01:27:14 +08:00
Terrence
6a7a403117 修复C3唤醒提示音 2025-07-21 14:40:35 +08:00
Terrence
a35a344f42 Fix frame samples for server AEC 2025-07-20 07:53:52 +08:00
Xiaoxia
3c71558a5f v1.8.0: Audio 代码重构与低功耗优化 (#943) 
* Reconstruct Audio Code

* Remove old IoT implementation

* Add MQTT-UDP documentation

* OTA升级失败时，可以继续使用
 2025-07-19 22:45:22 +08:00
laride
5c8707075f feat: add state change events and callbacks (#798) 2025-07-18 01:35:31 +08:00
ZhouKe
1aea59a472 multinet wakeword (#927) 
Co-authored-by: zk <982145@qq.com>
 2025-07-18 01:18:40 +08:00
Ky1eYang
0c971f76e3 feat: 添加自定义消息接收并打印在display或串口上 (#921) 
* feat: 添加声波配网, 需调整application的ReadAudio公有, 需添加条件编译, 位于'afsk_demod.h'内定义参数

* mod: afsk的重构，旨在提高代码可读性并遵循Google C++代码风格指南

* mod: 更新依赖esp-wifi-connect需求版号

* feat: 添加声波配网, 需调整application的ReadAudio公有, 需添加条件编译, 位于'afsk_demod.h'内定义参数

* mod: afsk的重构，旨在提高代码可读性并遵循Google C++代码风格指南

* mod: 更新依赖esp-wifi-connect需求版号

* mod: 添加判断只有在WiFi配置模式下才会调用ReadAudio, 否则delay（联网成功重启后该任务不会被启动）

* add: 添加USE_ACOUSTIC_WIFI_PROVISIONING进MENU开关声波配网功能

* feat: 添加mqtt自定义消息接收, 修复menu的声波配网config

* 简化接收custom消息的处理, 只打印或处理payload字段

* fix: 修复payload在执行时指向被释放的线程不安全问题

---------

Co-authored-by: yangkaiyue <yangkaiyue1@tenclass.com>
 2025-07-18 01:07:57 +08:00
Xiaoxia
6bb95073a4 v1.7.7: 支持 EC801E 模组，增加 xmini-c3-v3, xmini-c3-4g (#930) 2025-07-17 20:18:21 +08:00
Terrence
d1c586239c fix CPU usage of esp32c3 with ml307 2025-06-29 06:01:41 +08:00
Terrence
b3ab3d0920 在S3芯片上使用更多的PSRAM，解决立创开发板拍照时可能出现内存不足的问题 2025-06-24 04:59:00 +08:00
Terrence
bf125446b3 feat: Use BOOT button to enter audio testing state when Wi-Fi configuring 2025-06-13 19:57:07 +08:00
Terrence
7bc5f7bb0c Add audio debugger 2025-06-06 20:24:00 +08:00
Terrence
24ede22197 wechat ui支持摄像头预览 2025-06-06 16:54:53 +08:00
Xiaoxia
8729fad9a8 v1.7.1: fix send audio (#771) 
* screen -> display

* Add AUDIO_CODEC_DEFAULT_MIC_GAIN

* WebRTC -> NSNet

* fix: send audio priority
 2025-06-06 14:03:46 +08:00
Terrence
5621fa5a65 fix: 长按切换到点按后，无法唤醒的问题 2025-06-05 18:41:23 +08:00
Xiaoxia
ae57131c15 Add wake word to xmini-c3 (#730) 
* esp-hi: MCP protocol is not ready yet

* Add wake word to xmini-c3
 2025-05-31 22:21:03 +08:00
espressif2022
f5c1c30c5e feat: add new board esp-hi (#666) 
* feat: add new board esp-hi

* feat(esp-hi): update servo_dog_ctrl

---------

Co-authored-by: Li Junru <lijunru@espressif.com>
Co-authored-by: Xiaoxia <terrence@tenclass.com>
 2025-05-29 23:29:33 +08:00
Xiaoxia
d80f94387a v1.6.6: Set MCP as default IoT Protocol (#690) 2025-05-27 14:58:49 +08:00
Terrence
3a2e6875b6 Add audio_send_queue to synchronize timestamp 2025-05-27 05:44:46 +08:00
Terrence
0c57df1cd8 v1.6.5: Improve performance and memory usage 2025-05-26 14:30:44 +08:00
Terrence
0204b8800b update status bar immediately after network started 2025-05-24 07:25:34 +08:00
Terrence
0b98392286 Remove update display timer 2025-05-24 03:03:33 +08:00
Terrence
5da66773d5 Add MCP server 2025-05-22 19:19:48 +08:00
Terrence
7925f77739 Fix: 切换状态后一段时间内无法录音问题 2025-05-19 01:29:39 +08:00
Ky1eYang
27d85356e4 修改send packet时直接获取时间戳为队列获取时间戳 (#628) 
* 修改实时aec传输时间戳改队列方式

* 删/注释化一下log代码, pr

---------

Co-authored-by: yangkaiyue <yangkaiyue1@tenclass.com>
 2025-05-16 18:46:20 +08:00
Xiaoxia
be18c1aa04 Add server AEC option 2025-05-09 14:00:26 +08:00
Terrence
f6334246ca Add dummy audio processor 2025-04-29 18:17:08 +08:00
Xiaoxia
b804343d30 Audio stream packet with timestamp 2025-04-28 23:10:24 +08:00
Terrence
6fbf272613 Allow 10 audio packets in recv queue 2025-04-28 11:19:28 +08:00
Terrence
3404180a77 更新Wi-Fi组件版本，从OTA接口读取Websocket服务器 2025-04-21 15:13:11 +08:00
Terrence
c380617cca Support MQTT endpoint port 2025-04-21 15:13:11 +08:00
Terrence
6e73a11ac9 Add two types, system and alert. 2025-04-14 16:02:41 +08:00
Terrence
d5d8b34b2b Add activation version 2 2025-04-14 15:44:06 +08:00
Xiaoxia
f76f31aa12 Prevent too many opus packets in queue 2025-04-13 23:12:44 +08:00
Terrence
78d7b40914 将项目版本更新至1.5.9，优化内存使用，增加主任务栈大小至8192，优化固件升级流程，重构主循环为MainEventLoop，添加新版本检查功能，更新音频编解码器的DMA配置常量。 2025-04-13 11:47:15 +08:00
Xiaoxia
04c0da059f Fix missing display 2025-04-11 03:11:32 +08:00
Terrence
a5dfe67504 move ota headers to ota.cc 2025-04-10 18:56:17 +08:00
Xiaoxia
70883058e8 fix startup sound 2025-04-04 15:29:44 +08:00
Terrence
45012e38d4 Fix upgrade failure (main task stack overflow) 2025-03-30 16:13:49 +08:00