Files
agent-estate-wiki/queries/retained-research-answers.md
T

6.7 KiB
Raw Blame History

title, type, status, created, updated, verified_on, review_after, confidence, tags, sources
title type status created updated verified_on review_after confidence tags sources
Retained Research Answers query active 2026-07-22 2026-11-04 2026-11-04 2027-01-12 medium
query
research
vnc
remote-desktop
web-research
turboVNC-benchmark-pdf
superuser-15337
tigervnc-manpage
session-20261011

Retained Research Answers

Research answers worth keeping: what was asked, the answer, the evidence it rests on, and what the answer does not establish. Entries are point-in-time; re-verify anything load-bearing before acting on it.

The 2026-07-22 stub was filled on 2026-10-12 with the first retained answer (asked in the 2026-10-11 interactive session talaria-ios, 41 messages).

2026-11-04 re-verification: the answer's framing was wrong — it was written for "TigerVNC + Xtigervnc servers" that the estate does not run. The client-side advice remains valid for the one VNC server that does exist (x11vnc on CT460). See the correction note under the answer.

Fastest VNC client (asked 2026-10-11)

Answer for the CT460 VNC server: use the TigerVNC native viewer with AutoSelect and the defaults — free, and the fastest standard-RFB client.

⚠ Corrected 2026-11-04 — the original question was framed against servers the estate does not run. This answer was written on 2026-10-12 for "the CT460/CT200 TigerVNC + Xtigervnc servers". Neither exists. Live state, re-verified 2026-11-04:

  • CT460 runs x11vnc, not TigerVNC and not Xtigervnc: x11vnc -display :99 -forever -shared -rfbport 5901 -rfbportv6 5901 -noxdamage -quiet -nopw (PID 1290295), fronted by websockify → noVNC on 127.0.0.1:6080 → 5900 (a dead target — no listener on 5900) and 127.0.0.1:6081 → 5901 (live). The display itself is Xvfb :99 -screen 0 1600x1200x24 -ac -nolisten tcp.
  • CT200 runs no VNC server at all. Its only X server is Xvfb :495478600 -screen 0 800x680x24, started 2026-10-30 — that is the Camofox browser's own headless Xvfb, and the Camofox VNC plugin tree lives on CT460 (/home/hermes/services/camofox-browser/plugins/vnc/), not on CT200. No x11vnc, no Xtigervnc, no vncserver, and nothing listening on 5900/5901/6080/6081.

The advice itself still stands, because x11vnc speaks standard RFB and TigerVNC's viewer is the right client for it — but it now applies to one host, and the page's "the estate's own VNC/Xtigervnc setup" pointer is wrong. The real setup is documented at systems/headless-server, which already described x11vnc correctly back on 2026-10-20; this query page was the outlier. Lowest bandwidth is KasmVNC (WebP, ~30 % less) but that is a server swap, not a client. RealVNC Connect is not a faster decoder, it is a forked protocol (RFB 6, audio, no port-forwarding) and needs a RealVNC server. TurboVNC remains the reference for lossless 3D/CAD. Genuinely low latency means leaving RFB entirely: Parsec or Sunshine + Moonlight.

The load-bearing finding: the server-side encoder is the bottleneck, not the client. A viewer can only decode what it is sent. Two independent sources say the same thing — the TurboVNC benchmark report (encoder throughput measurements) and the accepted Super User answer (RFB ships pixel rectangles, so cost scales with resolution and bit depth).

Practical levers, server side (in order of payoff):

  1. Compression level 1–2. The single biggest lever; the 2012 report found compression got worse above level 6 while costing 4.4× the CPU.
  2. 16-bit colour where banding is acceptable — halves the pixel payload.
  3. Local cursor rendering (-UseLocalCursor) — large perceived win on slow links.
  4. Let AutoSelect work; forcing Tight on a fast link or Raw on a slow link just wastes CPU/bandwidth.
  5. Tunnel over SSH — latency and loss usually dominate codec choice.
  6. -FrameRate (~5) is worth setting over the internet; if the requested rate exceeds what the link delivers you get visible catch-up lag.

Client side: -CompressLevel defaults to 2 and is a client parameter (0 = low, 9 = high) — do not confuse it with the server-side encoder setting from the 2012 report. TigerVNC's viewer has an in-tree H.264 decoder only when built with ENABLE_H264=ON (FFmpeg or WMF backend); server-side H.264 support is still the blocker. Decode speed depends heavily on libjpeg-turbo SIMD — client CPU matters as much as client software (a thin client with no SIMD headroom decodes slowly regardless of viewer). noVNC is convenience, not speed — the JS/Canvas path is pure overhead versus the native client.

Evidence / sources:

  • TurboVNC "TigerVNC vs TurboVNC" benchmark, primary source (2012 data on 2012 CPUs): https://turbovnc.org/pmwiki/uploads/About/turbototiger.pdf — after the TigerVNC 1.2 encoder rework the two matched within ~2 % at compression level 1, and TigerVNC led by up to 16 % on indexed-colour workloads.
  • Super User question 15337 (recovered via the Stack Exchange API, no browser needed): RFB transmits pixel rectangles while RDP/X send drawing instructions — which is why VNC only wins at very low bitrates while remaining the most portable option.
  • TigerVNC man page, pulled from GitHub master because tigervnc.com is dead (domain parked at Gandi): -CompressLevel default 2, client-side.
  • RealVNC performance / viewer-comparison pages (recovered via Camofox): the "4.7× faster video / 4.5× faster CAD" claim is RealVNC against the previous RealVNC at 37 Mbps / 33 ms / 0.3 % loss — not a comparison against TigerVNC or TurboVNC. Their differentiator is proprietary RFB 6.
  • TigerVNC issue #1187 — H.264 server support status.

Does NOT establish: exact throughput multipliers on current hardware (the benchmark is 2012); no end-to-end benchmark was run against Anthony's own hosts. If speed becomes a real complaint, measure the existing server before changing clients. It also does not establish anything about TigerVNC or Xtigervnc server-side behaviour on this estate, because the estate runs neither — x11vnc's own encoder settings (-compress, -quality, -encodings) are the live lever on CT460, not the TigerVNC/TurboVNC encoder options the report discusses. The compression-level and local-cursor levers above map onto x11vnc as -compress <level> and the x11vnc cursor-shape options, but the specific numbers were tuned for a different server and should be re-measured before being trusted on this one.