1.9 KiB
Real VT1200 boot recording — reference material
Source: video/DEC VT1200 Hardware X11 terminal boot and demo. [j_mJB3asDgE].mkv
(201 MB, gitignored — it is a YouTube download and trivially re-obtainable).
Everything committed here is derived from it: contact sheets for navigation, full-resolution frames for detail.
Regenerating
V="video/DEC VT1200 Hardware X11 terminal boot and demo. [j_mJB3asDgE].mkv"
# contact sheets: one frame per 2 s, 6x6 tiles
ffmpeg -v error -y -i "$V" -vf "fps=1/2,scale=320:-1,tile=6x6" sheet_%02d.png
# full-resolution frame at a chosen time — OUTPUT seek; input seek (-ss before -i)
# fails on this filename with "Invalid argument"
ffmpeg -v error -y -i "$V" -ss 26 -frames:v 1 real_t26.png
The real boot timeline
This is the ground truth the harness is compared against.
| video time | what is on screen |
|---|---|
| 0:00–0:12 | nothing — the monitor's own blue no-signal backdrop |
| ~0:13 | the terminal starts driving the display |
| 0:14–0:44 | the self-test pie / progress disc sweeping to completion |
| ~0:45 | wipe to black |
| ~0:48 | a dialog box (the Terminal Manager) |
| ~1:08 | a light (mostly white) screen |
The harness reaches the X11 weave and the Terminal Manager via
VT1200_SKIP_LEVEL1=1; on real hardware level 1 passes, so it goes further.
One correction kept deliberately
The low-resolution contact sheets appeared to show a text block at ~0:14, and
that was briefly reported as a real screen. A full-resolution crop
(real_text_crop.png, taken at 0:14.2) shows the screen is blank blue — the
monitor's no-signal state. The "text" was a scaling/compression artefact of a
320-pixel tile.
A contact sheet tells you when something happens, not what it says. Crop and up-scale before reading anything off one — the same class of mistake as concluding "it never paints" from a single frame.