Keep PCM and its clock sample in cancellable datagrams so recovery
cannot replay audio buffered before a Wi-Fi stall.
Drop blocked real-time packets with an explicit discontinuity and make
a newer reconnect supersede stale audio authorities.
Isolate input sends from shared transport accounting and reserve
reliable capacity.
Prioritize QUIC execution and bound stale motion through the Windows
pipe and HID queues.
Route input through a dedicated high-priority stream, send pointer
motion as cumulative datagrams, and recover suspected loss reliably.
Reduce downstream input queuing with age-aware coalescing and input
thread scheduling.
Keep a stable target per transport binding and fan playback and state
notifications out to every active source. Serialize capture and clock
authority while preserving exact BUSY retries and failure cleanup.
A video configuration epoch can advance while client feedback for the
previous epoch is still in flight. Treat structurally valid stale video
control packets as obsolete instead of terminating the QUIC session.
Preserve a partially transmitted frame while replacing queued stale
frames. Continuous capture can otherwise discard every frame before the
client receives its final fragment.
Ignore runtime-selected alignment when comparing frame textures.
This allows software-adapter products through publication and routing.
Log the first execution-stage failure per graph.
Replace the single-owner PCM queues with one shared LGMP SPMC stream.
Give each admitted client reliable state and an independent playback
cursor, while clients without clock authority resample locally.
Extend the audio protocol with per-client epochs and barriers.
Keep capture exclusive and bound authority backpressure so slow
secondary readers cannot affect the disciplined low-latency path.
Give microphone data priority over replaceable clock feedback and
reduce feedback traffic on the shared client-to-host FIFO.
Retry short FIFO stalls outside the realtime thread and coalesce
consecutive packets in the existing fixed Windows buffers.
Avoid one broker write per tiny PipeWire quantum without adding
allocations or intentional capture latency.
