Media died silently whenever a client's source address changed. The relay bound a
peer's endpoint once and refused to move it, and the client stopped offering its
binding token after the first bind, so a Wi-Fi/cellular handover stranded the
session in both directions. Add an authenticated Rebind media frame: the binding
token travels in the clear for peer lookup only, and the AEAD tag over header and
token plus the peer's existing replay window are what authorize the move, so a
captured rebind cannot be replayed to redirect someone else's downlink. The client
rebuilds its UDP socket instead of retrying on one still pinned to a vanished
interface.
Nothing judged the control connection live: pings were sent and pongs ignored, so a
blackholed TCP path went unnoticed for minutes while the UI showed a live session.
Treat any server traffic as liveness and fail the connection when it stops, which
drives the existing reconnect.
The receive jitter buffer had lost its depth floor, so a channel without FEC or
DRED played out with no buffer at all and ordinary reordering became concealment.
Restore a one-frame floor, observe every arrival rather than only accepted ones —
a shallow buffer was rejecting the late arrivals that should have deepened it —
and allow playout to hold a frame so depth can follow a degrading link. A stalled
consumer now sheds the oldest queued packet instead of refusing the live talkspurt.
Add a deterministic network-impairment simulation covering bursty loss, jitter,
reordering, duplication, outages and a stalled consumer, a handover test against a
real relay, a replay test for the rebind path, and a blackholed control connection
driven through a freezable TCP proxy.
Declare XSAppIconAssets in the app manifest so actool receives --app-icon
and the bundle carries CFBundleIconName, and always pass both Xcode version
placeholders to the ReplayKit extension so an empty build number cannot
drop CFBundleVersion and fail installd.
Proposal for replacing the C++ core, C++ server, and Swift macOS/iOS
clients with a single .NET 10 / C# codebase.
Covers the dependency map (8 vcpkg deps + 1 vendored -> 3 native libs),
the real-time-audio design, per-client strategy, a test-porting plan for
all 29 ctest cases, doc-sync work, an 11-phase migration, and a risk
register.
Two findings drive the shape of the plan:
- SslStream has no RFC 5705 keying-material exporter, which the media
AEAD key derivation depends on (docs/security.md 2). The API is an
unapproved proposal and SChannel structurally cannot export secrets.
Recommends BouncyCastle's managed TLS 1.3 stack, which does implement
the exporter and keeps the wire format byte-compatible with the C++
implementation -- so the existing tree stays usable as a conformance
oracle throughout the port. Protocol-v3 in-band media keys documented
as the fallback.
- The iOS ReplayKit broadcast upload extension stays in Swift: 50 MB
jetsam cap plus an unsupported extension type in .NET for iOS, and it
already doesn't link the core. Leaves one Swift file plus the shared
App Group ring.
Indexed in docs/README.md. Nothing here is implemented yet.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds vcpkg as a submodule at vcpkg/, pinned to the exact commit vcpkg.json
already declares as builtin-baseline, so the bundled checkout and the
manifest's resolved port versions can never drift apart.
cmake/voicecat-toolchain.cmake, scripts/common.sh, and
clients/apple/scripts/build-xcframework.sh now resolve vcpkg as:
VCPKG_ROOT env var (external checkout) > bundled submodule. Docs updated
to describe the new one-time setup.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Drop the M0 no-deps skeleton preset and all VOICECAT_HAS_NET/AUDIO/OPUS/NS
guards that it required. Every subsystem is fully implemented; the stub
#else paths were dead code that added noise to every header and source file.
- CMakePresets.json: remove skeleton configure/build/test entries
- CMakeLists.txt (root/core/tests): remove VOICECAT_USE_VCPKG_DEPS option
and guards; all targets now build unconditionally
- 17 C++ source files: unwrap HAS_* guards, delete stub #else blocks
- apm_processor.cpp: delete ApmPassthrough no-op class; create() always
returns RnnoiseProcessor
- 18 test files: remove HAS_* guards and stub int main() skip bodies
- docs/building.md: remove skeleton from preset table and prose
VOICECAT_HAS_LOOPBACK (Windows WASAPI loopback platform gate) unchanged.
29/29 ctest green.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Three bugs fixed across the full stack (proto/server/core/ABI/Win/macOS/iOS):
1. Join/Leave Voice now truly subscribes/unsubscribes from the voice plane.
Previously the button only toggled the local mic — receiving was always on
(gated by channel membership alone). Added a protocol-level voice subscription
concept: new SubscribeVoiceRequest/UnsubscribeVoiceRequest/VoiceSubscriptionResult
proto messages, User.voice_subscribed field, vc_join_voice/vc_leave_voice C ABI
functions, VC_EVENT_VOICE_STATE event, server-side voice_subscribed flag checked
by the SFU relay recipient filter, and core-client gating of remote-stream
decoder setup. All three clients rewired to subscribe+mic on Join / unsubscribe
on Leave. Text chat works regardless of voice subscription.
2. Channel edit dialog now shows the channel's actual current settings. The read
struct vc_channel was missing sort_order and audio fields — only the write
struct vc_channel_info had them. Extended vc_channel with both (additive, no
ABI break), updated the session model and list_channels marshaling to populate
them, and updated all three clients' edit callers to use actual channel info
instead of hardcoded defaults.
3. Channel parameter updates now automatically restart everyone's streams.
Previously editing a channel's audio config persisted and broadcast a
ChannelEvent::UPDATED, but no layer restarted streams — encoders/decoders are
frozen at announce time. handle_channel_event now detects audio-config changes
on the user's current channel and stop->starts each active local stream. The
server reads the updated config on re-announce; peers wire up fresh decoders
at the new ssrc.
All 29 CTest tests pass; Windows DLL + C# client build clean. Apple clients not
yet compile-verified (Windows environment).
The per-listener noise-reduction toggle (vc_set_remote_stream) did nothing
on Windows/macOS/iOS. The decode loop gated the RNNoise pass on
dec_channels == 1 as a proxy for "this stream is voice" (assuming
stereo => screen-share). The stereo-mic capture commit broke that: a stereo
mic with send-side NR off transmits stereo Opus, so the receiver decoded
two channels and skipped NR entirely. gain/mute have no channel guard, which
is why only NR appeared broken.
Thread the stream kind through init_recv_stream into RemoteStream::is_voice
(set from si.kind() == STREAM_MIC), gate receive NR on is_voice instead of
channel count, and fold a stereo voice frame to mono -> denoise -> duplicate
back across both channels in place (symmetric with the send-side downmix;
RNNoise is mono-only). Screen-audio shares are never denoised.
New test test_recv_noise_reduction drives AudioEngine and asserts a stereo
voice stream's noise floor collapses with NR on (RMS 1046 -> 0.1) while a
screen-audio share stays unchanged. ctest --preset dev green 29/29.
Docs: voice.md section 10. Shared-core fix; clients need only a rebuild.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Both desktop mics were hard-mono: the core defaults capture_channels=1 and
neither client ever called vc_set_capture_channels (only iOS did). Add a
persisted "Stereo microphone" toggle to each client's Audio settings, applied
when the mic stream starts and live via vc_set_capture_channels + vc_audio_restart.
Expose both ABI calls in the Windows interop; the macOS wrapper already had them.
Core fix: encode_and_send_frame now folds a stereo mic frame to mono on a mono
channel - previously the channels==2 branch encoded interleaved L/R directly even
on a mono channel, feeding a mono opus_encode 2x its samples (wrong pitch/garbage).
Real stereo still only reaches the wire on a stereo channel; on a mono channel the
mic is cleanly downmixed.
Test: test_stereo_mic_mono_channel. ctest --preset dev green (28/28). Docs: voice.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The two-sided NR plumbing (RemoteStream::recv_ns + the per-listener
vc_set_remote_stream noise_reduction toggle) was wired but inert:
ApmProcessor::create() returned a no-op passthrough, because the
originally-planned webrtc-audio-processing has no working Windows/macOS
build. Drop in RNNoise as the real backend behind the same ApmProcessor
interface, lighting up both NR paths.
- Vendor RNNoise (BSD-3 + CC0) at third_party/rnnoise/ — the vcpkg port
is !windows !arm, so it can't cover our primary targets. Shrunk int8
model (78MB -> 11.7MB via upstream scripts/shrink_model.sh), built as a
standalone C static lib with no RTCD (portable scalar path on x86,
auto-NEON on arm64) under -DDISABLE_DEBUG_FLOAT. Model is baked in
(rnnoise_create(NULL)); no runtime file.
- New RnnoiseProcessor (core/src/audio/apm_processor.cpp) selected by
ApmProcessor::create() when VOICECAT_HAS_NS. Mono/48kHz/480-sample;
our clock is fixed 48kHz and Opus frame sizes are multiples of 480, so
no resampling. RT-safe: allocates at construction, lock-free in the
capture/playback callbacks.
- Receive-side: lit up via the factory; gated to mono streams (a stereo
stream is a screen-audio share, not voice).
- Send-side (new): vc_set_input_noise_reduction(client, enable) ABI +
vc_client::mic_ns_, run before input gain/VAD in on_capture_frame. A
stereo mic is downmixed to mono ONLY when NR is on — with NR off a
stereo mic keeps full stereo (never collapse mic quality unasked).
- Enable C as a project language for the vendored lib.
- New noise_suppression test: white noise through ApmProcessor::create()
drops ~99.9% RMS. ctest --preset dev green, 28/28. windows-client DLL
builds clean with vc_set_input_noise_reduction exported, system-only deps.
- Docs synced: voice.md §10, tech-stack.md §1/§5, third_party/README.md,
vcpkg.json note, PROGRESS.md, CLAUDE.md.
Client on/off UI toggles (Windows/macOS/iOS) are the remaining follow-up.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The iOS audio path was a hybrid: Voice-Chat-class presets ran a native
VPIO AVAudioEngine (core external) while Stereo/Studio/A2DP presets ran
the core's miniaudio devices. Nearly every "no input / no output / both"
bug lived in the seam between the two paths — the lingering miniaudio
capture unit fighting VPIO, the audioRestart ordering dance, the
route-change "glitching" loop, stereo<->mono stickiness, and
"can't hear anyone". Switching presets/routes mid-call routinely dropped
a direction.
Drive ALL iOS audio through one AVAudioEngine with the core fully
external at all times: setExternalPlayback(1) once at connect, every MIC
stream external_feed=1, mic via vc_stream_feed_pcm, playback via
vc_set_mixed_output_sink (drained by an always-on AVAudioSourceNode so
remote audio plays before joining voice). VPIO + AGC toggle per preset.
Every preset/route/interruption change funnels through one deterministic
Swift-only reconfigure (stop -> apply session config -> rebuild -> start)
— no second path to hand off to, so a change can't drop a direction.
- IOSVoiceProcessingEngine.swift -> IOSAudioEngine: always-on source-node
playback, conditional mic tap, VPIO/AGC; one rebuild() backing
startListening/stop/startMic/stopMic/reconfigure/setCaptureChannels.
- IOSAudioRouter: 7 presets -> 4 (Voice Chat / Stereo Mic / Mono Mic /
Advanced); persisted voiceProcessingEnabled + agcEnabled; setters call
IOSAudioEngine.reconfigure() instead of audioRestart/reconcileVoicePath.
- AudioSessionManager slimmed; SessionState mic lifecycle collapsed;
AppState wires external playback + listening at connect, stop at
disconnect; SettingsView shows 4 presets + Advanced VPIO/AGC toggles.
No core/ABI/test changes — relies on the already-shipped external API
(test_external_pcm, test_external_playback). xcodebuild iOS device Debug
BUILD SUCCEEDED. Updates docs/voice.md §8 and PROGRESS.md.
The encoder set OPUS_SET_INBAND_FEC, but the decoder never invoked FEC --
the loss path went DRED -> PLC, so FEC redundancy was emitted (and paid for
in bitrate) yet never consumed.
Wire FEC recovery into AudioEngine::on_playback between DRED and PLC: copy
the next buffered packet once, try DRED, else (if the stream negotiated FEC)
decode(next_pkt, ..., fec=true), else PLC. Recovery priority is now
DRED -> FEC -> PLC. Add per-stream RemoteStream::fec_enabled_, captured from
OpusParams in init_recv_stream. Docs (voice.md) updated to match.
ctest --preset dev: 27/27 green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Latency between speakers grew to multiple seconds and only reset on
rejoining voice. Root cause was the receiver playout logic, not the
codec settings: the playout clock free-ran in real time while the
sender omitted silence from its timestamps (and set no header flags),
and the only correction snapped the clock to the *oldest* buffered
frame — which could only ever add standing latency. target_depth_ms_
was computed but never enforced, so latency could only grow or reset.
Fix: bound playout against the stream's leading edge (newest frame).
(Re)seed to the leading edge on start/marker/starve (no prebuffer, so
latency stays low), and frame-skip catch-up trims any backlog beyond
target+hysteresis — the missing downward force.
Hardening: sender now stamps kFlagMarker (talkspurt start) and kFlagDtx,
consumed on recv for clean resync; adaptive late-drop window; EWMA
outlier rejection so silence gaps/stragglers don't poison the estimate;
duplicate counting and ring-underrun diagnostics.
New test_jitter_depth asserts depth stays bounded (<200ms) while
arrivals outrun playback. ctest --preset dev green (27/27).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The per-channel sample_rate field was inert after pinning the codec to
48 kHz. Make it meaningful without changing the 48 kHz clock: carry it as
OpusParams::max_bandwidth_hz and apply OPUS_SET_MAX_BANDWIDTH in
OpusEncoder::init (8000->narrowband, 16000->wideband, 24000->super-wideband,
48000->full). A low-bitrate room can now shed out-of-band content while
every endpoint keeps a single 48 kHz clock.
Make sample_rate channel-authoritative on the server: conn_session no
longer overrides effective sample_rate with the client's always-48000
request (it now behaves like frame_ms/mode). vc_get_stream_audio_config
reports the channel's configured rate for own streams too.
New ctest channel_samplerate: a 7 kHz tone is attenuated ~1000x on an
8 kHz (narrowband) channel vs a 48 kHz (full-band) channel, proving the cap
is in effect. ctest --preset dev 26/26.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The AudioEngine capture clock is fixed at 48 kHz / 20 ms (960-sample
frames), but a channel may set any Opus frame_ms (2.5..60 ms, voice.md
§3) and the server enforces it unclamped. on_capture_frame handed the
engine's 960-sample frame straight to an encoder configured for the
channel's window: frame_ms > 20 was silently ignored, and frame_ms < 20
broke entirely (receiver sized its decode buffer too small ->
OPUS_BUFFER_TOO_SMALL -> dead audio). Affected the hardware mic and
vc_stream_feed_pcm alike.
Reframe each captured/fed block to ls.frame_samples via a per-LocalStream
accumulator (pre-sized at announce, no RT-thread alloc) before
encode_and_send_frame; the 20 ms case stays a zero-copy fast path. Also
pin the codec to 48 kHz in opus_params_from_audio_config — it was honoring
a non-48k effective sample_rate against a 48 kHz PCM clock.
New ctest frame_ms_reframe covers 40 ms (accumulate) and 10 ms (split)
feed->encode->relay->decode->sink round trips. ctest --preset dev 25/25.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
"All apps except selected" previously captured the complement of a frozen
app snapshot in INCLUDE mode (missed late-launched apps and system sounds,
wasted captures on silent windows). It now opens a single ProcessLoopbackCapture
in EXCLUDE mode (AUDIOCLIENT_PROCESS_LOOPBACK_MODE_EXCLUDE_TARGET_PROCESS_TREE)
of the one chosen app — true system-mix-minus-one, dynamic so apps launched
after sharing starts are included. The picker enforces single-selection in
exclude mode (the activation params take one target PID).
Adds an "Exclude VoiceCat's own audio (prevents echo)" checkbox (default on,
entire-desktop only) that routes the desktop capture through the same EXCLUDE
path targeting our own process id, killing the whole-device self-echo loop.
No C++/ABI changes. Updates voice.md and PROGRESS.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
iOS "voice chat" had echo and no noise suppression: real iOS AEC/NS/AGC
come only from Apple's Voice-Processing I/O unit (VPIO), but the core
plays/captures via miniaudio's plain RemoteIO units, so .voiceChat mode
alone never engaged AEC.
Core (ABI PATCH 1->2):
- vc_set_mixed_output_sink + vc_set_external_playback. In external mode the
AudioEngine opens no hardware playback device; a mixer-timer thread drives
on_playback (decode+mix) on a ~20ms cadence and ships the final mix to the
sink. start() also skips the hardware capture device when the MIC stream is
external_feed (AudioParams.external_capture).
- New white-box test test_external_playback (drives the timer with no hw).
iOS/Swift:
- StreamDescriptor.externalFeed; VoiceCatClient.setMixedOutputSink /
setExternalPlayback wrappers.
- IOSVoiceProcessingEngine: AVAudioEngine + setVoiceProcessingEnabled; mic
tap -> feedPcm, mixed-sink lock-free ring -> AVAudioSourceNode (both share
the VPIO unit so AEC has its reference signal).
- IOSAudioRouter.currentConfigUsesVoiceProcessing scopes VPIO to the AEC
presets; SessionState join/leave + reconcileVoicePath() switch paths;
Voice Chat defaults to speaker; Settings surfaces AEC/NS state.
Known: pending on-device verification; a few bugs to fix afterward.
A bad UDP packet on a flaky link could permanently wedge the voice path,
unrecoverable even across app restarts. Three defects:
1. Anti-replay window was advanced from the UNAUTHENTICATED header seq
before the AEAD tag was checked, and not rolled back on failure. One
corrupted/forged frame shoved recv_highest_ far ahead, after which every
legitimate frame was rejected as "too old" forever. Reorder to
replay-check -> authenticate -> update (RFC 3711 3.3); the window now
moves only after a successful tag check.
2. The wire seq was only the low 16 bits of the nonce counter (zero-extended
on receive). After 65,536 frames the nonce desynced and all frames failed
auth. Widen the voice frame seq u16 -> u64 (header 14 -> 20 bytes). The
core owns all UDP framing, so Swift/C# clients need only a rebuild. This
is a versioned wire change: VOICECAT_PROTOCOL_VERSION 1 -> 2, handshake
rejects on mismatch.
3. Server leaked per-session UDP state on disconnect; unregister_session now
frees udp_endpoints_/udp_tokens_/ssrc_to_session_.
Also add rate-limited dropped-frame logging to MediaRelay so a wedged media
path is observable. New regression tests in test_media_aead.cpp cover the
poison (fails on old code) and the 16-bit wrap. ctest --preset dev
-E external_pcm: 22/22 pass (external_pcm aborts on a pre-existing CoreAudio
shutdown race, unrelated).
Let users choose what the SCREEN_AUDIO stream captures before sharing:
share everything, only selected apps, or all except selected apps, plus a
first-class "Exclude screen reader (VoiceOver) audio" toggle.
ScreenCaptureKit filters audio per application, so ScreenAudioCapture now
takes a ScreenAudioSelection and builds the matching SCContentFilter
(including:/excludingApplications:). New ScreenSharePickerSheet lists
running apps from SCShareableContent. iOS left untouched -- ReplayKit only
delivers the mixed system stream, so per-app filtering is impossible there.
Implement system/desktop audio sharing on the Apple clients, feeding the
existing SCREEN_AUDIO Opus -> AEAD -> UDP path via vc_stream_feed_pcm. No
C++/protocol/codec changes -- the core was already ready (the Windows-only
loopback is #ifdef VOICECAT_HAS_LOOPBACK; off Windows the stream just waits
for fed PCM). Audio only; video is dropped.
macOS (in-process):
- ScreenAudioCapture.swift drives an audio-only SCStream
(excludesCurrentProcessAudio), converts Float32 -> int16 in the channel's
mono/stereo mode, and calls feedPcm. Capture starts on the self
.streamStarted event (effective config known then). Wired into
MainWindowController.screenAudioClicked().
iOS (forward-to-host, single session):
- VoiceCatBroadcast: a ReplayKit Broadcast Upload Extension consumes
.audioApp only, resamples to 48kHz int16 stereo (AVAudioConverter), and
writes a shared App Group SPSC ring (BroadcastAudioRing.swift). It does
not link libvoicecat.
- Host BroadcastAudioPump drains the ring (reacting to the extension's
Darwin notifications) and feeds the SCREEN_AUDIO stream it owns, downmixing
to mono when the channel is mono. Screen audio appears as a second stream
of the same user; no credentials persisted. UI is RPSystemBroadcastPicker
View in VoiceControlsView. Removes the speculative BroadcastCredentials.
Docs: voice.md s9, CLAUDE.md status, PROGRESS.md.
Talon
