Files
chanora/crates/chanora_audio/Cargo.toml
T

170 lines
7.7 KiB
TOML

[package]
name = "chanora_audio"
description = "Chanora audio subsystem — platform-native capture/playback, audiopus encode, tsclientlib AudioHandler for decode + jitter buffer + mix. DEC-011, DEC-011.1."
version.workspace = true
edition.workspace = true
rust-version.workspace = true
authors.workspace = true
license.workspace = true
repository.workspace = true
publish.workspace = true
[dependencies]
chanora_protocol = { path = "../chanora_protocol" }
thiserror.workspace = true
tracing.workspace = true
sonora = "0.1"
webrtc-vad = "0.4"
# ndarray is required by ort's tensor construction API and by
# Silero VAD ONNX inference across all platforms.
ndarray = "0.17"
# Opus encoder. tsclientlib already pulls this; we depend explicitly so
# this crate can compile against it without going through tsclientlib.
audiopus = "0.3.0-rc.0"
# AudioHandler lives in the tsclientlib crate behind the `audio`
# feature. We import the crate just for the AudioHandler type; the
# Connection type stays inside chanora_protocol.
tsclientlib = { git = "https://github.com/ReSpeak/tsclientlib.git", rev = "04aa2491", default-features = false, features = ["audio"] }
tokio = { version = "1", features = ["sync", "rt", "macros", "time"] }
rustfft = "6.2.0"
[target.'cfg(all(not(target_os = "android"), not(target_os = "ios"), not(target_os = "macos")))'.dependencies]
# Desktop audio I/O for Windows capture/playback and Linux capture.
# Linux playback uses SDL2; Apple platforms use direct VoiceProcessingIO
# AudioUnits via `coreaudio-rs` for the voice path.
cpal = "0.17.3"
[target.'cfg(any(target_os = "ios", target_os = "macos"))'.dependencies]
# Direct CoreAudio AudioUnit access on Apple platforms (DEC-011 follow-up).
# cpal's Apple path does not expose the voice-processing controls Chanora
# needs for VoIP. Use `kAudioUnitSubType_VoiceProcessingIO` directly via
# `coreaudio-rs` so capture/playback share Apple's native AEC + AGC + NS
# voice unit on both iOS and macOS.
#
# Default features keep `audio_toolbox` + `core_audio`, both required
# for AudioUnit construction + property access.
coreaudio-rs = "0.14"
# Grand Central Dispatch bindings — used to run AudioUnit initialize/start
# on the main queue to avoid the VPIO RPC timeout on iOS simulator.
dispatch2 = "0.3"
[target.'cfg(target_os = "ios")'.dependencies]
# ONNX Runtime Rust binding for Silero VAD v6 (P1 VAD_002). The official
# iOS CocoaPod ships ONNX Runtime as a static framework, so iOS links it
# into chanora_bridge at build time instead of loading a dylib at runtime.
ort = { version = "2.0.0-rc.12", default-features = false, features = ["std", "ndarray"] }
[target.'cfg(not(target_os = "ios"))'.dependencies]
ort = { version = "2.0.0-rc.12", default-features = false, features = ["load-dynamic", "ndarray", "api-24"] }
[target.'cfg(target_os = "android")'.dependencies]
# JNI bindings to flip Android's AudioManager into MODE_IN_COMMUNICATION
# when the voice-comm preset is requested. ndk_context is initialised
# by the bridge crate's android_init shim.
jni = { version = "0.21", default-features = false }
ndk-context = "0.1"
# Oboe-rs (Google Oboe wrapper) for low-latency voice capture + playback.
# Primary backend for SDD-111..SDD-115. The pre-compiled static library
# shipped with `oboe-sys` 0.6 covers armv7 / aarch64 / x86 / x86_64.
# Default features keep the precompiled library + pregenerated bindings
# so we avoid the clang-sys / libclang requirement on the build host.
#
# Using local fork edisonjwa/oboe-rs (v0.6.2) with:
# - catch_unwind safety in callbacks
# - unwrap_or_default in enum getters (no more SessionId panic)
# - get_raw_session_id() for JNI hardware effect binding
# - deduplicated macro impls
# - PowerSavingOffloaded PerformanceMode variant
oboe = { git = "https://github.com/EdisonJwa/oboe-rs", rev = "a14f9b83ecea8c93f5a692f2ee7808445b938c35" }
[target.'cfg(target_os = "windows")'.dependencies]
# Real Windows global PTT (SDD-083 / SDD-084): RegisterRawInputDevices
# + WM_INPUT translation backed by a hidden message-only window, and
# SetWindowsHookExW(WH_KEYBOARD_LL / WH_MOUSE_LL) fallback. Both
# require a per-backend OS thread that owns a message pump.
windows = { version = "0.54", features = [
"Win32_Foundation",
"Win32_Graphics_Gdi",
"Win32_System_LibraryLoader",
"Win32_System_Threading",
"Win32_UI_Input",
"Win32_UI_Input_KeyboardAndMouse",
"Win32_UI_WindowsAndMessaging",
] }
[dev-dependencies]
# `test-util` enables `start_paused` / virtual-clock tests used by
# the missed-key-up watchdog unit tests.
tokio = { version = "1", features = ["sync", "rt", "macros", "time", "test-util"] }
# Cross-platform recording Layer for the SDD-090 / DEC-027 privacy
# invariant integration test (`tests/ptt_privacy.rs`).
tracing-subscriber = { version = "0.3", features = ["registry"] }
# SDD-120 §3 — criterion bench harness (realtime_capture / opus_codec /
# resampler). `harness = false` per bench entry below disables the
# default libtest harness so criterion can install its own.
criterion = "0.5"
# SDD-120 §3 item 1 — dhat is used as the global allocator inside
# `benches/realtime_capture.rs` to count post-warmup heap allocations
# on the realtime capture path. Dev-dep only — does NOT affect
# production builds.
dhat = "0.3"
# SDD-120 §5 / §8 — JSON serialization for `emit_baseline` /
# `compare_baseline` binaries that consume criterion's per-bench
# `estimates.json` outputs and emit the SRS-217 baseline schema.
serde_json = "1"
[[bench]]
name = "realtime_capture"
harness = false
path = "benches/realtime_capture.rs"
[[bench]]
name = "opus_codec"
harness = false
path = "benches/opus_codec.rs"
[[bench]]
name = "resampler"
harness = false
path = "benches/resampler.rs"
[target.'cfg(target_os = "linux")'.dependencies]
# GNOME-on-Wayland Global Push-to-Talk uses the freedesktop
# `org.freedesktop.portal.GlobalShortcuts` interface over D-Bus.
# `zbus` is the standard async D-Bus crate; the `tokio` runtime
# selector is mandatory in zbus 5; we share the tokio runtime
# the rest of the audio + core crates already depend on. The
# `blocking-api` feature is retained so the audio-engine
# probe path can do a synchronous portal-version read without
# starting an async runtime; the live session flow uses the
# async surface.
zbus = { version = "5", default-features = false, features = ["tokio", "blocking-api"] }
# Stream / sink utilities for consuming portal signals on the
# async path.
futures-util = { version = "0.3", default-features = false, features = ["std"] }
# Random token bytes for the portal handle_token / session_handle_token
# options. The portal recommends fresh tokens to scope its own
# object paths per call.
rand = "0.8"
# SDL2 audio for Linux. Replaces the cpal playback path on Linux only;
# cpal stays in use for Linux capture and Windows capture/playback.
# Apple platforms use direct VoiceProcessingIO AudioUnits. Rationale: the
# cpal Linux backend opens raw ALSA `default`, which on most Arch
# / Fedora / Debian installs routes through `dmix` + `plug` with
# nearest-neighbour resampling and very small period sizes — the
# combination produces audible crackling/popping. SDL2 on the same
# systems routes through PipeWire's PulseAudio compat bridge (or
# real PulseAudio), both of which carry a high-quality resampler
# and a sensible default period. The upstream tsclientlib audio
# example (`tsclientlib/examples/audio_utils/ts_to_audio.rs`) and
# the official Qint client both use SDL2 in exactly this shape;
# this dep brings Chanora in line with that pattern.
#
# `bundled` is OFF deliberately — we link against the system
# libSDL2.so. Arch ships `sdl2-compat`; Debian/Ubuntu ship
# `libsdl2-2.0-0`; Fedora ships `SDL2`. The chanora-flutter Linux
# build documentation lists this as a runtime dependency.
sdl2 = { version = "0.37", default-features = false }