impress_voice_input/CMakeLists.txt
impressionyang 01a39ddc8c feat: 实现基于 VAD 的流式 WAV 录制识别方案
- 新增 StreamingAudioWriter 组件:持续录制 WAV 文件,VAD 检测静音段自动切换
- 静音检测:检测到 ~1s 连续静音后关闭当前文件,触发 chunkCompleted 信号
- STTTestPage 重构:移除缓冲区推理模式,改为 WAV 文件流式识别
- 每个 WAV 文件完成后在后台线程读取并推理,不阻塞继续录制
- 设置页面新增「调试音频目录」配置项
- 音频存储路径:debug 模式使用配置目录,非 debug 模式使用系统临时目录

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-05-13 15:33:03 +08:00

195 lines
5.7 KiB
CMake
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

cmake_minimum_required(VERSION 3.20)
project(impress_voice_input
VERSION 0.1.0
LANGUAGES CXX
DESCRIPTION "基于 ONNX 的实时语音转文本输入法"
)
set(CMAKE_CXX_STANDARD 17)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_EXTENSIONS OFF)
# 默认使用 Release 带调试信息
if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
set(CMAKE_BUILD_TYPE RelWithDebInfo CACHE STRING "Build type" FORCE)
set_property(CACHE CMAKE_BUILD_TYPE PROPERTY STRINGS
"Debug" "Release" "RelWithDebInfo" "MinSizeRel")
endif()
set(CMAKE_AUTOMOC ON)
set(CMAKE_AUTORCC ON)
set(CMAKE_AUTOUIC ON)
set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
# ============================================================================
# 依赖查找
# ============================================================================
list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/cmake")
# Qt 6
find_package(Qt6 REQUIRED COMPONENTS Core Widgets Concurrent Network)
# Windows 不需要 D-BusLinux/macOS 按需加载
if(WIN32)
find_package(Qt6 REQUIRED COMPONENTS Gui)
else()
find_package(Qt6 REQUIRED COMPONENTS DBus)
endif()
qt_standard_project_setup()
# ONNX Runtime
include(dependencies)
# ============================================================================
# 源文件
# ============================================================================
set(SOURCES
src/main.cpp
# App
src/app/application.cpp
src/app/config_manager.cpp
# Core (平台无关)
src/core/stt_engine.cpp
src/core/sense_voice_engine.cpp
src/core/sense_voice_features.cpp
src/core/sense_voice_tokenizer.cpp
src/core/mel_spectrogram.cpp
src/core/whisper_tokenizer.cpp
src/core/audio_processor.cpp
src/core/vad.cpp
src/core/voice_input_service.cpp
# Audio
src/audio/audio_capture.cpp
src/audio/audio_decoder.cpp
src/audio/audio_ring_buffer.cpp
src/audio/streaming_audio_writer.cpp
# UI
src/ui/main_window.cpp
src/ui/stt_test_page.cpp
src/ui/file_transcribe_page.cpp
src/ui/settings_page.cpp
src/ui/widgets/audio_waveform.cpp
src/ui/widgets/text_output.cpp
src/ui/widgets/progress_panel.cpp
src/ui/widgets/hotkey_recorder.cpp
# Utils
src/utils/logger.cpp
src/utils/timer.cpp
src/utils/string_utils.cpp
)
set(HEADERS
src/app/application.h
src/app/config_manager.h
src/core/stt_engine.h
src/core/sense_voice_engine.h
src/core/sense_voice_features.h
src/core/sense_voice_tokenizer.h
src/core/sense_voice_cmvn.h
src/core/mel_spectrogram.h
src/core/whisper_tokenizer.h
src/core/audio_processor.h
src/core/vad.h
src/core/voice_input_service.h
src/audio/audio_capture.h
src/audio/audio_decoder.h
src/audio/audio_ring_buffer.h
src/audio/streaming_audio_writer.h
src/ui/main_window.h
src/ui/stt_test_page.h
src/ui/file_transcribe_page.h
src/ui/settings_page.h
src/ui/widgets/audio_waveform.h
src/ui/widgets/text_output.h
src/ui/widgets/progress_panel.h
src/ui/widgets/hotkey_recorder.h
src/utils/logger.h
src/utils/timer.h
src/utils/string_utils.h
)
# ============================================================================
# 平台特定源文件
# ============================================================================
if(WIN32)
# Windows 实现RegisterHotKey + SendInput
list(APPEND SOURCES src/core/win_hotkey.cpp src/core/win_text_injector.cpp)
list(APPEND HEADERS src/core/win_hotkey.h src/core/win_text_injector.h)
add_compile_definitions(PLATFORM_WINDOWS)
elseif(APPLE)
# macOS 实现待完善Carbon EventManager / CGEvent
list(APPEND SOURCES src/core/mac_hotkey.cpp src/core/mac_text_injector.cpp)
list(APPEND HEADERS src/core/mac_hotkey.h src/core/mac_text_injector.h)
add_compile_definitions(PLATFORM_MACOS)
else()
# Linux 实现D-Bus Portal + XTest
list(APPEND SOURCES src/core/caps_lock_voice_hotkey.cpp src/core/wayland_text_injector.cpp)
list(APPEND HEADERS src/core/caps_lock_voice_hotkey.h src/core/wayland_text_injector.h)
add_compile_definitions(PLATFORM_LINUX)
endif()
add_executable(${PROJECT_NAME} ${SOURCES} ${HEADERS})
target_include_directories(${PROJECT_NAME} PRIVATE
${CMAKE_CURRENT_SOURCE_DIR}/src
${CMAKE_CURRENT_SOURCE_DIR}/third_party
${ONNXRUNTIME_INCLUDE_DIRS}
${PORTAUDIO_INCLUDE_DIRS}
${DR_LIBS_INCLUDE_DIR}
${NLOHMANN_JSON_INCLUDE_DIR}
)
target_link_libraries(${PROJECT_NAME} PRIVATE
Qt6::Core
Qt6::Widgets
Qt6::Concurrent
Qt6::Network
$<$<BOOL:${WIN32}>:Qt6::Gui>
$<$<NOT:$<BOOL:${WIN32}>>:Qt6::DBus>
${ONNXRUNTIME_LIBRARIES}
${PORTAUDIO_LIBRARIES}
$<$<NOT:$<BOOL:${WIN32}>>:pthread>
)
target_compile_options(${PROJECT_NAME} PRIVATE
$<$<CXX_COMPILER_ID:GNU,Clang>:-Wall -Wextra -Wpedantic>
$<$<CXX_COMPILER_ID:MSVC>:/W4>
)
# ============================================================================
# 资源文件
# ============================================================================
# 样式表
qt_add_resources(${PROJECT_NAME} "styles"
PREFIX "/"
FILES
src/ui/resources/styles/main.qss
)
# ============================================================================
# 安装
# ============================================================================
install(TARGETS ${PROJECT_NAME}
RUNTIME DESTINATION bin
BUNDLE DESTINATION .
)
# ============================================================================
# 测试
# ============================================================================
option(BUILD_TESTS "构建单元测试" OFF)
if(BUILD_TESTS)
enable_testing()
add_subdirectory(tests)
endif()