- Add Devices.Equals() method to compare USB device configurations
- Add GetGadgetDevices() to retrieve current device state
- Skip gadget reconfiguration when device state is unchanged
- Remove 6 unused audio translation keys from all language files
- Add config fields for bitrate, complexity, DTX, FEC, buffer periods
- Add RPC methods for get/set audio config and restart
- Add UI settings page with controls for all audio parameters
- Add Apply Settings button to restart audio with new config
- Add config migration for backwards compatibility
- Add translations for all 9 languages
- Clean up redundant comments and optimize log levels
The previous default EDID did not advertise audio capabilities,
preventing HDMI audio capture from working. This update ensures
the JetKVM properly identifies itself to source devices and
enables audio capture out of the box.
Changes:
- Display identifies as "JetKVM HDMI" (manufacturer ID: JTK)
- Includes full audio support (2/6/8-channel LPCM, up to 192 kHz)
- Advertises all TC358743XBG capabilities (1080p60, YCbCr, Deep Color)
This allows HDMI audio to work by default without manual EDID
configuration.
The continuous ensure_sleep_mode_disabled() call in the format detection
loop caused repeated I2C transactions to the TC358743, disrupting HDMI
audio capture. Sleep mode is already disabled once during video_init(),
which is sufficient.
- Revert .golangci.yml to dev branch state (removed custom build-tags)
- Remove internal/audio/cgo_source_stub.go (not needed with proper cross-compilation)
- Fix import ordering in ui/src/utils.ts
Use 'make lint-go' for proper ARM cross-compilation environment.
Provides no-op AudioSource implementations for platforms that don't
support ARM CGO audio (x86_64, darwin, etc.). This allows golangci-lint
to run successfully on any platform without requiring ARM cross-compilation
toolchain.
The stub implementations return errors when called, ensuring that if
they're accidentally used at runtime on non-ARM platforms, it will fail
gracefully with a clear error message rather than undefined symbols.
Build constraints ensure the real CGO implementation is used on linux/arm
and linux/arm64, while stubs are used everywhere else.
Remove all subprocess-based audio code to simplify the audio system and
reduce complexity. Audio now uses CGO in-process mode exclusively.
Changes:
- Remove subprocess mode: Deleted Supervisor, IPCSource, embed.go
- Remove audio mode selection from UI (Settings → Audio)
- Remove audio mode from backend config (AudioMode field)
- Remove JSON-RPC handlers: getAudioMode/setAudioMode
- Remove Makefile targets: build_audio_output/input/binaries
- Remove standalone C binaries: jetkvm_audio_{input,output}.c
- Remove IPC protocol implementation: ipc_protocol.{c,h}
- Remove unused IPC functions from audio_common.{c,h}
- Simplify audio.go: startAudio() instead of startAudioSubprocesses()
- Update all function calls and comments to remove subprocess references
- Add constants to cgo_source.go (ipcMaxFrameSize, ipcMsgTypeOpus)
- Keep update_opus_encoder_params() for potential future runtime config
Benefits:
- Simpler codebase: -1,734 lines of code
- Better performance: No IPC overhead on embedded hardware
- Easier maintenance: Single audio implementation
- Smaller binary: No embedded audio subprocess binaries
The audio system now works exclusively via CGO direct C function calls,
with ALSA device selection (HDMI vs USB) still configurable via settings.
Audio quality improvements:
- Enable constrained VBR to prevent bitrate starvation at low volumes
- Increase Opus complexity from 2 to 5 for better quality
- Enable DTX for bandwidth optimization
- Enable FEC (Forward Error Correction)
- Add DTX and FEC signaling in SDP (usedtx=1;useinbandfec=1)
Default configuration changes:
- Change default audio output source from HDMI to USB
- Enable USB Audio device by default
- USB audio works on current stable image (HDMI requires newer device tree)
These changes fix crackling issues at low volumes and provide better
overall audio quality for both USB and HDMI audio paths.
Remove dynamic gain code and rely on Opus encoder quality improvements:
- Increase Opus complexity from 2 to 5 for better quality
- Change bandwidth from FULLBAND (20kHz) to SUPERWIDEBAND (16kHz) for better quality at 128kbps
- Disable FEC to allocate all bits to audio quality
- Increase ALSA buffer from 40ms to 80ms for stability
The dynamic gain code was adding complexity without solving the underlying
issue: TC358743 HDMI chip captures digital audio at whatever volume the
source outputs. Users should adjust volume at the source or in their browser.
Add noise gate threshold at peak > 256 (-42dB) to prevent dynamic gain
from amplifying quantization noise and hardware noise floor. This fixes
crackling, buzzing, and static-like noise when HDMI audio is at very
low volume or during silence.
Without the gate, signals below -42dB (peak < 256) would get 8x gain
applied, amplifying noise floor to audible levels. Now these signals
pass through unmodified, eliminating the artifacts.
- Check SetReadDeadline error in IPC client
- Explicitly ignore Kill() error (process may be dead)
- Remove init() function and rely on explicit ExtractEmbeddedBinaries() call
* feat: release keyPress automatically
* send keepalive when pressing the key
* remove logging
* clean up logging
* chore: use unreliable channel to send keepalive events
* chore: use ordered unreliable channel for pointer events
* chore: adjust auto release key interval
* chore: update logging for kbdAutoReleaseLock
* chore: update comment for KEEPALIVE_INTERVAL
* fix: should cancelAutorelease when pressed is true
* fix: handshake won't happen if webrtc reconnects
* chore: add trace log for writeWithTimeout
* chore: add timeout for KeypressReport
* chore: use the proper key to send release command
* refactor: simplify HID RPC keyboard input handling and improve key state management
- Updated `handleHidRPCKeyboardInput` to return errors directly instead of keys down state.
- Refactored `rpcKeyboardReport` and `rpcKeypressReport` to return errors instead of states.
- Introduced a queue for managing key down state updates in the `Session` struct to prevent input handling stalls.
- Adjusted the `UpdateKeysDown` method to handle state changes more efficiently.
- Removed unnecessary logging and commented-out code for clarity.
* refactor: enhance keyboard auto-release functionality and key state management
* fix: correct Windows default auto-repeat delay comment from 1ms to 1s
* refactor: send keypress as early as possible
* refactor: replace console.warn with console.info for HID RPC channel events
* refactor: remove unused NewKeypressKeepAliveMessage function from HID RPC
* fix: handle error in key release process and log warnings
* fix: log warning on keypress report failure
* fix: update auto-release keyboard interval to 225
* refactor: enhance keep-alive handling and jitter compensation in HID RPC
- Implemented staleness guard to ignore outdated keep-alive packets.
- Added jitter compensation logic to adjust timer extensions based on packet arrival times.
- Introduced new methods for managing keep-alive state and reset functionality in the Session struct.
- Updated auto-release delay mechanism to use dynamic durations based on keep-alive timing.
- Adjusted keep-alive interval in the UI to improve responsiveness.
* gofmt
* clean up code
* chore: use dynamic duration for scheduleAutoRelease
* Use harcoded timer reset value for now
* fix: prevent nil pointer dereference when stopping timers in Close method
* refactor: remove nil check for kbdAutoReleaseTimers in DelayAutoReleaseWithDuration
* refactor: optimize dependencies in useHidRpc hooks
* refactor: streamline keep-alive timer management in useKeyboard hook
* refactor: clarify comments in useKeyboard hook for resetKeyboardState function
* refactor: reduce keysDownStateQueueSize
* refactor: close and reset keysDownStateQueue in newSession function
* chore: resolve conflicts
* resolve conflicts
---------
Co-authored-by: Adam Shiervani <adam.shiervani@gmail.com>
Implement SIMD-optimized audio operations using ARM NEON for Cortex-A7 targets
Update Makefile and CI configuration to support NEON compilation flags
Add SIMD implementations for common audio operations including:
- Sample clearing and interleaving
- Volume scaling and format conversion
- Channel manipulation and balance adjustment
- Endianness swapping and prefetching
- Added check to not attempt auto update if time sync is needed and not yet successful (delays 30 second to recheck).
- Added resync of time when DHCP or link state changes if online
- Added conditional* fallback from configured* NTP servers to the IP-named NTP servers, and then to the DNS named ones if that fails
- Added conditional* fallback from the configured* HTTP servers to the default DNS named ones.
- Uses the configuration* option for how many queries to run in parallel
- Added known static IPs for time servers (in case DNS resolution isn't up yet)
- Added time.cloudflare.com to fall-back NTP servers
- Added fallback to NTP via hostnames
- Logs the resultant time (and mode)
Alex P
Aveline
Adam Shiervani
Marc Brooks
Alex
Marc