Files
punktfunk/docs-site/content/docs/roadmap.md
T
enricobuehler 0e977817f9
ci / docs-site (push) Successful in 58s
ci / rust (push) Failing after 1m3s
ci / web (push) Successful in 1m4s
apple / swift (push) Successful in 1m22s
decky / build-publish (push) Successful in 19s
docker / build-push (., web/Dockerfile, punktfunk-web) (push) Successful in 10s
docker / build-push (ci, ci/fedora-rpm.Dockerfile, punktfunk-fedora-rpm) (push) Successful in 9s
docker / build-push (ci, ci/rust-ci-noble.Dockerfile, punktfunk-rust-ci-noble) (push) Successful in 8s
docker / build-push (ci, ci/rust-ci.Dockerfile, punktfunk-rust-ci) (push) Successful in 8s
docker / build-push (docs-site, docs-site/Dockerfile, punktfunk-docs) (push) Successful in 46s
windows-host / package (push) Failing after 5m38s
ci / bench (push) Successful in 6m12s
docker / build-push (--build-arg FEDORA_VERSION=44, ci, ci/fedora-rpm.Dockerfile, punktfunk-fedora44-rpm) (push) Successful in 6m38s
apple / screenshots (push) Successful in 6m25s
docker / deploy-docs (push) Successful in 29s
deb / build-publish-host (push) Failing after 7m30s
deb / build-publish (push) Successful in 11m22s
rpm / build-publish (43, bazzite, punktfunk-fedora-rpm) (push) Failing after 8m59s
arch / build-publish (push) Successful in 11m57s
android / android (push) Successful in 15m26s
rpm / build-publish (44, fedora-44, punktfunk-fedora44-rpm) (push) Successful in 13m46s
feat(hdr): GNOME 50 HDR screencast capture + Linux Main10 encode
GNOME 50 (Mutter MR 4928, PipeWire >= 1.6) added HDR screen sharing for
monitor streams: 10-bit PQ formats (xRGB_210LE/xBGR_210LE) with MANDATORY
BT.2020 + SMPTE-2084 colorimetry props, advertised while the mirrored
monitor is in BT.2100 colour mode. Wire the Linux host into it end-to-end
on the GameStream desktop-mirror path (PUNKTFUNK_VIDEO_SOURCE=portal):

* pf-frame: PixelFormat::X2Rgb10/X2Bgr10 (DRM XR30/XB30; X2Bgr10 is the
  Windows Rgb10a2 layout) + fourccs.
* pf-capture: want_hdr portal offer — HDR-only LINEAR-dmabuf pods with
  MANDATORY PQ/BT.2020 props (SHM excluded: Mutter's SHM record path
  paints 8-bit ARGB32 regardless of format; tiled excluded: the EGL
  de-tile blit is 8-bit RGBA8), negotiated-colorimetry parse, generic
  HDR10 hdr_meta(), packed-10-bit CPU cursor blend, a process-wide SDR
  downgrade latch on negotiation timeout, and a DisplayConfig BT.2100
  colour-mode probe (gnome_hdr_monitor_active).
* pf-encode: libav NVENC X2RGB10->P010 swscale (BT.2020 limited) ->
  HEVC Main10 / 10-bit AV1 with PQ VUI; VAAPI 10-bit on both paths (CPU
  P010 upload + dmabuf XR30 scale_vaapi p010/bt2020); can_encode_10bit
  now probes for real on Linux; 10-bit sessions route around the
  8-bit-only Vulkan-video/direct-NVENC backends.
* GameStream: host_hdr_capable() Linux arm, live monitor-HDR check at
  RTSP honor time, capturer-pool reuse keyed on HDR-ness, gs_bit_depth
  covers the new formats. New `punktfunk-host hdr-probe` diagnostic and
  a PUNKTFUNK_SPIKE_HDR spike lever.
* Native plane stays honestly 8-bit via capturer_supports_hdr(): Mutter
  RecordVirtual streams are SDR-only upstream (GNOME 50 and 51-dev), so
  virtual-display sources cannot deliver HDR yet.

Validated on the RTX 5070 Ti (GNOME 50.3 / PipeWire 1.6.8): the Main10
probes pass and the ignored nvenc_hdr10_smoke GPU test emits an IDR that
ffprobe reads as Main 10 / yuv420p10le / bt2020nc / smpte2084 / limited.
Live HDR capture negotiation still needs an HDR monitor on glass; VAAPI
10-bit needs the AMD box.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 09:30:55 +02:00

7.3 KiB
Raw Blame History

title, description
title description
Roadmap What's shipped, what's in progress, and what's next for punktfunk.

A quick map of where punktfunk is today and where it's heading. For the detailed, dated changelog, see Status & Progress.

Legend: shipped · 🟡 in progress · 🔭 planned · blocked upstream

At a glance

Area
Protocol core — FEC · crypto · C ABI
GameStream host (works with Moonlight)
Native punktfunk/1 protocol
Linux host (KWin · GNOME · gamescope · Sway)
Windows host (NVIDIA · AMD · Intel) beta
Apple client (macOS · iOS · iPadOS · tvOS)
Linux client (GTK4 shell + Vulkan session)
Android client (phone · TV)
Windows client 🟡
Web console + pairing
Concurrent sessions (shared desktop)
Network speed test + bitrate
HDR / 10-bit streaming Windows host · 🚧 Linux host (GNOME 50+ desktop mirror; virtual displays blocked upstream)
Surround audio (5.1 / 7.1)
Sub-frame pipelining (latency) 🔭

Shipped

  • The host, two ways. The lower-latency native punktfunk/1 protocol (QUIC control + UDP data with GF(2¹⁶) Leopard FEC + AES-GCM) — the secure default — and, opt-in via serve --gamestream, a GameStream host any Moonlight client can use. Both run from one process.
  • Native-resolution virtual displays on Linux across KWin, GNOME/Mutter, gamescope, and Sway/wlroots, with a fully zero-copy GPU path to NVENC (stable 240 fps at 5120×1440).
  • A native Windows host (x64; NVIDIA/AMD/Intel encode) — a signed installer with secure-desktop capture and a bundled virtual-display driver, and the only host that can stream HDR (10-bit BT.2020 PQ, captured from an HDR Windows desktop and encoded as HEVC Main10). See Windows Host. (Beta — newer than the Linux host.)
  • Clients on every platform — native apps for Apple (macOS, iOS, iPadOS, tvOS), Linux, Android (phone + TV), and Windows, each with hardware decode, controllers including DualSense, audio + mic, and automatic host discovery. See Clients.
  • Secure by default — SPAKE2 PIN pairing with pinned reconnects, one-click delegated approval from the web console, and mDNS LAN auto-discovery.
  • Tuned for latency — concurrent sessions (stream one desktop to several devices at once), mid-stream resolution renegotiation, a cross-machine clock-skew handshake, a 1 Gbps+ data plane, an in-app network speed test that informs the bitrate picker, and automatic adaptive bitrate (the encoder re-targets mid-stream when bitrate is set to Automatic).
  • Surround sound — 5.1 and 7.1 end to end: the host encodes multichannel via multistream Opus and the native clients render more than two channels, with clean, synchronous stereo where the path is stereo.
  • Clipboard sync — bidirectional text and images between host and client, carried on a side plane over the native protocol's QUIC channel.

🟡 In progress

  • Windows client on-glass validation. Hardware decode and the GUI are validated on NVIDIA and Intel; the HDR present path still needs verification on real HDR hardware.
  • Apple presenter polish. The lower-latency VTDecompressionSessionCAMetalLayer stage-2 path is now the default; HDR brightness and 4:4:4 still need on-glass validation.
  • Web console parity. Surfacing the speed test and bitrate picker the apps already have.
  • Windows host hardening. Broader real-world testing — especially on-glass validation of the AMD (AMF) and Intel (QSV) encode paths, which are CI-green but newer than NVENC.

🔭 Planned

  • Magic multi-user support. Map a connecting client to a real user account on the host and log them in automatically. The client picks an identity — e.g. an Apple TV profile — which maps to an available host profile (likely behind a second per-user auth layer); on connect to an idle host, the user lands in their own desktop/session, signed in, without touching the host. Turns one box into a personal, profile-aware streaming target for a household.
  • Object-based spatial audio. With 5.1/7.1 surround shipped, the frontier is object-based sound. Capturing game audio objects on native Windows is blocked by a closed renderer API, but Proton already routes them through Wine's ISpatialAudioClient — where it currently discards the dynamic/height objects and their 3D positions. Finishing that rendering would give every Proton gamer real spatial sound, and tunnelling the objects + positions to the client would let punktfunk spatialize them on the client — head-tracked remote spatial audio (to AirPods and the like) that no streaming stack does today.
  • Sub-frame pipelining. Overlap encode and transmit within a single frame (a direct NVENC slice path) — the next big latency lever at high resolutions.
  • True glass-to-glass latency measured end to end (capture → on-screen present).
  • gamescope multi-user isolation. Per-session input and audio so concurrent clients are fully independent desktops (the shared-desktop case already works).
  • Peer-approved pairing. Approve a new device from an already-paired device's own app.
  • WAN / anywhere access. Reach a host beyond the LAN — NAT traversal (ICE/STUN/TURN) plus a self-hostable relay for when no direct path can be punched, and QUIC connection migration so a client roams between networks (Wi-Fi ↔ cellular) without dropping the session. The control plane is already QUIC over UDP, so this is mostly signalling and relay rather than a protocol change.
  • VRR / adaptive-sync passthrough. Variable refresh end to end — the host renders at a variable rate and the client presents with tearing-control/VRR instead of a fixed cadence, for tear- and judder-free gaming. Builds on the client's presentation-feedback path and the per-session virtual outputs.
  • Desktop quality-of-life. More of the essentials that make remote work pleasant, each a new side plane over the existing QUIC datagram channel: multi-monitor streaming (present the host's several outputs as separate client windows) and virtual-webcam redirection (the client's camera shows up as a webcam on the host, so video calls run on the remote machine). (Clipboard sync already shipped — see above.)

Parked / blocked

  • HDR / 10-bit on the Linux host — virtual displays. GNOME 50 added HDR screencasting for monitor streams, and the host now uses it: the GameStream desktop mirror (PUNKTFUNK_VIDEO_SOURCE=portal) negotiates the 10-bit PQ formats and encodes HEVC Main10 BT.2020 PQ (punktfunk-host hdr-probe reports readiness; on-glass validation pending). What stays blocked upstream is HDR on virtual monitors — Mutter's RecordVirtual streams are still SDR-only (through the GNOME 51 dev branch), so the native protocol and GameStream's default virtual-display source stream 8-bit until that lands; the host is ready the moment it does.
  • Advanced DualSense voice-coil haptics. Scoped and shelved (it rides the controller's USB audio interface, with near-zero game support on Linux). Adaptive triggers, rumble, and the lightbar already ship.