Commit Graph
88 Commits
Author SHA1 Message Date
jimmyandClaude Opus 5.5 c71b6da9d9 Rearrange the test show's console and square off its colour buttons.
The layout, chase colours, mover speeds and levels as saved from QLC+: the
colour frames stacked on the left with the cans and movers pads beside them.
The Colour 1 and Colour 2 buttons are 75 px squares, and the Chase runs and
Pads layer frames below Colour 2 move down to clear them
(scripts/shows/square_colours.py). The Movers speed dial comes in from
x=1820 to beside Movers position, and both speed dials are trimmed to the
dial and its time.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 15:38:24 +13:00
jimmyandClaude Opus 5.5 c7546e5ea9 Show in the test show whether the pads are on the cans or the movers.
A Pads layer frame under the movers pads has Cans pads and Movers pads
buttons (scripts/shows/layer_indicator.py). The pad controller holds channel
16 note 0 (cans) or 1 (movers) down and the other up whenever Pause changes,
QLC+ reloads or it starts, and the Flash buttons light while their note is
held, so the lit one is the layer and a missed message can't leave them the
wrong way round. Each flashes an empty scene of its own: with one between
them, QLC+ lit both.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 15:09:49 +13:00
jimmyandClaude Opus 5.5 467b480c76 Put the pads on the cans or movers layer from Pause's value.
The SMC-PAD latches Pause itself: one message a press, 127 and 0 on
alternate presses, nothing on release. The pad controller flipped the layer
only on 127, so every second press did nothing. Bank 1 now follows the
value: 0 is the cans, 127 the movers.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 15:09:47 +13:00
jimmyandClaude Opus 5.5 31ccbab340 Make a click on a cans chase button run the whole chase in the test show.
Each chase was two buttons on one pad note: the pads frame's ran the side
dimmer hold under Peak, and the Chase runs frame's ran the stripe outside it.
The pad pressed both; a click pressed only the hold, which lit nothing. Now
the pads frame's button runs a hold of the chase's own whose step is the Sides
dimmer and a Script that starts the stripe. Peak reaches the dimmer through the
chaser and collection but not the stripe through the script, as with the pad
(scripts/shows/chase_clicks.py). The Chase runs buttons no longer listen to the
pad.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 13:57:27 +13:00
jimmyandClaude Opus 5.5 cf94c7b1ba Press the colours again when a restarted QLC+ links itself to the pad controller.
The pad controller only took a QLC+ reload as one when it had to make the
link to QLC+ itself. A restarted QLC+ often subscribes to the virtual port on
its own as it opens the show, before the next check, so the colours stayed off.
Each check now also notes the QLC+ process and port it feeds (qlc_feeds), and a
new one counts as a reload too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 13:57:22 +13:00
jimmyandClaude Opus 5.5 4fd11445e8 Add cabaret's can patterns to the test show.
Eight of make_cabaret_vc.py's Can Patterns on the cans pads 1-8, in a Can
patterns solo frame in place of the pads frame's empty Buttons 53-60:
Theater, Rainbow, Odd Even, Left Right, Scanner (with a fading tail), Fill,
Center and Pairs, laid out for the test show's cans and walking the ring round
the room (scripts/shows/can_patterns.py). Collapse and Chase Out are left out.

The masks show the Colour 1 colour; Odd Even (a checkerboard) and Left Right
swap between Colour 1 and Colour 2 through four Matrix widgets in the Chase
colours frame. All of them follow the cans' Speed dial (knob 1). Each pattern is
RGB matrices running from start to stop (rgbscripts/steps.js and swap.js, linked
into ~/.qlcplus/rgbscripts), so a chase started over one lands on top, and the
pad controller restarts the pattern that is on just after a Colour 1 press so
the new colour scene doesn't cover it.

The row of 4 (#3-#6) never animates; CLAUDE.md now says so for whatever comes
next.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 13:57:13 +13:00
jimmyandClaude Opus 5.5 e7e1cf656b Remove the sound-reactive cans from the test show.
The Sound and Threshold audio triggers go, with the Sound submaster beside the
cans' Front and Back sliders (CC 100) and the Sound peak frame with its Peak
submaster and Hit button and the Cans hit functions only Hit ran. Pads 1 and 2
are plain spare buttons again (scripts/shows/drop_sound.py). The pads frame's
own Peak (knob 6) stays: it is the chases' peak.

The pad controller drops its side of it: the Pump and Threshold modes, the
CC 100 resets, holding the side knobs at the peak and the pad lamps that showed
which mode was on. Pads 1 and 2 and knobs 5, 6 and 8 pass straight through.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 13:57:00 +13:00
jimmyandClaude Opus 5.5 3b38df3dd0 Smooth the movers' Corners in the test show.
Corners slammed the heads about 100° of pan every 0.6 s at full motor speed.
It now runs at a quarter of the speed dial like the other shapes, about 1.4 s
a corner, and its pad starts a collection with a Movers glide scene that sets
the heads' Pan/Tilt Speed to 40 while it runs, so they ease into each corner.
That channel is forced HTP on the movers, so it drops back to full speed when
Corners stops instead of QLC+ holding it for everything after
(scripts/shows/mover_corners_glide.py).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 01:47:26 +13:00
jimmyandClaude Opus 5.5 f674dc9fe0 Give the movers every EFX shape and drop their Strobe in the test show.
The movers layer loses its Strobe (pad 9): its button, its scene and the
forced-HTP strobe channel it needed. Its pad, the rest of the second row
and pad 5 take the five shapes the movers did not have yet: Diamond, Square,
Corners (SquareChoppy), Leaf, and Tilt (Line2 on tilt alone). All nine sit in
the Movement solo frame, one at a time, mirrored and relative to the knobs.
Knob 13's speed dial gives Corners an eighth of its time and the rest a
quarter (scripts/shows/mover_all_efx.py). The pad controller settles the
knobs after all nine.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 01:47:15 +13:00
jimmyandClaude Opus 5.5 c3fe965e79 Add Circle and Sweep for the movers to the test show.
They fill the movers layer's top row, pads 15 and 16 (notes 98 and 99), in
the Movement solo frame with Wander and Ballyhoo: a circle of about ±45° of
pan and ±21° of tilt, and a straight pan sweep of about ±60°, both mirrored
and relative to the knobs. Knob 13's speed dial gives them a quarter of its
time, 5.5 s a round to start with (scripts/shows/mover_more_efx.py). The pad
controller settles the knobs after them too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 01:47:05 +13:00
jimmyandClaude Opus 5.5 80df550916 Run the movers' Wander and Ballyhoo one at a time.
Their two buttons go into a Movement solo frame where they sit, so starting
one stops the other (scripts/shows/mover_movement_solo.py). The pad
controller follows: it tracks the one that is on, and settles the knobs only
when that one goes off. Swapping needs no settle, as the new EFX writes over
the old one's offset.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 01:46:45 +13:00
jimmyandClaude Opus 5.5 09aab70bd6 Put the movers' Wander and Ballyhoo speed on knob 13 in the test show.
A Movers speed dial on knob 13 (CC 42), beside the cans' Speed dial, sets
how long Wander takes to go round and gives Ballyhoo an eighth of it
(scripts/shows/mover_speed.py).

QLC+ 4.12 counts an EFX's fade-in inside every round, so Wander has been
taking 22 s and Ballyhoo 2.5 s, not 24 s and 3 s. The dial keeps those as
its starting point and runs from 62 s and 7.5 s fully left to 10 s and 1 s
fully right. The fade-ins stay, so the movers don't jump when either starts.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 01:46:27 +13:00
jimmyandClaude Opus 5.5 ae186c51ea Drop the movers' chases and put Wander and Ballyhoo on the top row.
The movers layer loses its Chase, Chase reverse, Zigzag and Opposing pads,
with their runs, holds, colour matrices, fixture groups and Speed dial
entries. Wander and Ballyhoo move up to pads 13 and 14 (notes 96 and 97),
and the pad controller settles the knobs after those notes instead
(scripts/shows/mover_pads_top_row.py).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 01:46:18 +13:00
jimmyandClaude Opus 5.5 913494c0aa Add a Ballyhoo pad for the movers to the test show.
Pad 11 on the movers layer (note 94) toggles a fast figure of eight around
the Pan and Tilt knobs, like Wander: about ±85° of pan and ±26° of tilt, all
four movers together so the pairs mirror left to right and front to back
(scripts/shows/mover_ballyhoo.py).

The pad controller now settles the knobs after Ballyhoo goes off as well as
Wander, and keeps track of each separately.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 01:46:03 +13:00
jimmyandClaude Opus 5.5 8caf341d30 Point the back movers at the back wall at pan knob 0 in the test show.
M3 and M4 are mounted 90° round from M1 and M2, so at pan 0 they face the
middle. Knob 9 now drives a Pan slider for M1 and M2 over 0–212 and a Pan
back slider for M3 and M4 over 43–255 (scripts/shows/mover_pan_back.py).
The same span keeps the pairs 90° of pan apart all the way round: at knob 0
the front pair face the audience and the back pair the back wall, and the
left and right pairs point at the centre and the side walls together.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 00:49:37 +13:00
jimmyandClaude Opus 5.5 00c8c2d35d Add a Wander pad for the movers to the test show.
Pad 10 on the movers layer (note 93) toggles a relative EFX that drifts all
four movers around the Pan and Tilt knobs: a 24 s Lissajous of about ±30° pan
and ±10° tilt, each mover a quarter of the way further round, easing in over
2 s (scripts/shows/mover_wander.py).

QLC+ 4.12 leaves a relative EFX's last offset on the movers when it stops,
until the sliders under it change. So when Wander goes off, the pad
controller nudges the Pan and Tilt knobs a step and puts them back, 0.1 s
apart.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 00:49:33 +13:00
jimmyandClaude Opus 5.5 f8c8cea3ef Free the pad's ALSA clients when the Private port is missing.
On Bluetooth there is no Private port, so every Sound or Threshold press
opened two sequencer clients to light the lamps and left them open when the
port lookup raised. After 29 presses the sequencer was full and aconnect
could no longer open it. set_sound_lamps and Link now close both clients
on that path too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 00:49:25 +13:00
jimmyandClaude Opus 5.5 9c1ef6f2d4 Read the pad over Bluetooth first, and USB when it isn't paired.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 00:02:08 +13:00
jimmyandClaude Opus 5.5 2586ead4fa Put the chase speed on knob 1 in the test show.
A Speed dial on knob 1 (CC 30) sets the step time of the cans' and movers'
chase and zigzag runs, from 800 ms fully left to 50 ms fully right, starting
at the old 200 ms. Zigzag forward stays at half the step, and the dimmer holds
follow at 8 steps so they stretch with the run. The 500 ms trail is unchanged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 00:02:08 +13:00
jimmyandClaude Opus 5.5 fc8abf9c99 Set the movers' zeros and pan directions, swap M1/M2's patch, and model tilt in the simulator.
The front right mover is on DMX 1 and the front left on 45, so M1 and M2
swap addresses in the stage map and test.qxw. Measured on the rig, pan 0
faces forward on the front pair and into the middle on the back pair, back
at 85. M1 and M4 turn anticlockwise, so movers get a pan_reverse flag.

Tilt 127 now points straight down: the simulator draws a spot under the
head, and past 127 the beam flips behind the pan heading.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 22:54:14 +13:00
jimmyandClaude Opus 5.5 2b0f9e2f9b Re-arrange the test show's console and put the movers' brightness with their position knobs.
The Dimmers sliders are now Middle, Front and Back, and the movers' Brightness
(knob 15) leads the Movers position frame.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 20:47:22 +13:00
jimmyandClaude Opus 5.5 e96b31ab66 Re-aim cans 1, 2, 7 and 8 on the XCHC stage map.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 20:03:18 +13:00
jimmyandClaude Opus 5.5 2219788904 Give the test show Dimmers knobs instead of the custom colour mixer, and add Amber.
The colour scenes leave the dimmers to Level sliders on knobs 15, 5, 7 and 8.
The Chataigne export carries Level sliders and their starting values.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 20:03:18 +13:00
jimmyandClaude Opus 5.5 11da656ee6 Toggle the pad's bank 1 to the movers with Pause, and add a Threshold sound mode.
Pause flips bank 1 between the cans (36-51) and the movers (84-99). Threshold
flashes the cans on bass hits and is exclusive with Sound, and knob 6 is the
sides' peak. The Sound and Threshold pads light while on, and the colour pads
that were on are pressed again when QLC+ reloads the show. --lock-knobs is gone.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 20:03:18 +13:00
jimmyandClaude Opus 5.5 72c227b153 Merge dev into main: the pad controller, simulator, Chataigne export and test show.
Keep main's Temple of Techno show over dev's older copy from the first venue
workspaces commit.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 20:00:21 +13:00
jimmyandClaude Opus 5.5 86d87bf8cf Add a movers bank and Pan, Tilt, Focus and Rotate knobs to the test show.
Pause toggles bank 1 between the cans and the movers (scripts/shows/mover_bank.py).
Knobs 9-12 (CC 38-41) drive all four movers' pan, tilt, focus and lens
rotate, as in templeoftechno.qxw (scripts/shows/mover_position.py).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 19:56:38 +13:00
jimmyandClaude Opus 5.5 18f0ac2dc5 Add a virtual pad and an end-to-end test against the simulator, and run tests with pytest.
padcontroller.virtualpad presses pads and sets knobs through padcontroller-in,
refusing unless QLC+ last opened shows/test.qxw. tests/test_e2e_pad.py uses it
to check can and mover colours in the simulator's /api/state (PAD_E2E=1).

pytest is a dev package now, so the plain test functions unittest skipped run
too. test_small_eye.py is a rig tester, not a test, so pytest leaves it alone,
and the #14 direction assertion now matches the stage map (-179).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 17:58:04 +13:00
jimmyandClaude Opus 5.5 d522b1a2cf Add sound-reactive cans, a strobe pad and side chases to the test show.
The Sound pad (bank 1 note 36) toggles a QLC+ Audio Triggers widget whose
volume bar drives a Sound submaster beside the can dimmers, so the cans'
brightness follows the music while the movers stay steady. padcontroller sends
CC 100 at full on every press and release of that pad, bypassing its CC
smoothing, so switching sound off always leaves the cans at full.

test.qxw also gains Chase forward/reverse, a double and an opposing zigzag
down the side cans (RGB matrices whose colour follows the Colour 2 pads), a
Strobe flash pad with the cans' strobe channel forced HTP, and the Dimmers
frame, forced-LTP side colours and dimmer-free colour scenes from parallel
sessions.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 17:03:27 +13:00
jimmyandClaude Opus 5.5 6bbb3d5a29 Add CLAUDE.md: check the pad controller and simulator are up, and keep the test show on 127.0.0.1.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 14:34:05 +13:00
jimmyandClaude Opus 5.5 643f7c8fb6 Send the test show's Art-Net to 127.0.0.1 for the simulator.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 14:34:05 +13:00
jimmyandClaude Opus 5.5 8f8c69898a Read the pad over USB when it's plugged in, and keep its links to the pad and QLC+ up.
padcontroller.midi now follows the pad every 2 s: it reopens it after a
replug, moves to USB when that turns up, and re-links QLC+ to the virtual
port after a restart, so QLC+ can stay open. Only the chosen source is
subscribed, so a USB pad no longer gets Bluetooth input as well.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 14:34:05 +13:00
jimmyandClaude Opus 5.5 a295fefac0 Pick the simulator's stage map from a drop-down, and make the QLC+ show optional.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 14:05:42 +13:00
jimmyandClaude Opus 5.5 df05c8ca49 Add a systemd user service and installer for the stage simulator.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 13:29:23 +13:00
jimmyandClaude Opus 5.5 20f018a523 Let the padcontroller installer install the service without enabling it at login.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 13:27:48 +13:00
jimmyandClaude Opus 5.5 9742121144 Rename padeditor to padcontroller, with the MIDI intercept as padcontroller.midi and its virtual port as padcontroller.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 13:25:06 +13:00
jimmyandClaude Opus 5.5 1e09cb3825 Rename the MIDI intercept to padcontroller, as a module and as a service.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 12:33:43 +13:00
jimmyandClaude Opus 5.5 00de2595b6 Move the SMC-PAD MIDI intercept into padeditor and add a systemd user service with an installer.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 12:06:31 +13:00
jimmyandClaude Opus 5.5 647cfa3a75 Group scripts into pad/, shows/ and anim/ folders.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 11:54:36 +13:00
jimmyandClaude Opus 5.5 5f593b973c Move the SMC-PAD packet captures into captures/.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 11:46:01 +13:00
jimmyandClaude Sonnet 5.5 e007978ef9 Add SMC-PAD packet captures and a preview config for the pad editor.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-22 11:44:19 +13:00
jimmyandClaude Sonnet 5.5 e2392ecd0b Keep the simulator fitted to the window and label fixtures on hover.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-22 11:44:19 +13:00
jimmyandClaude Sonnet 5.5 729a315c7a Generate a Chataigne project from a show spec.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-22 11:44:19 +13:00
jimmyandClaude Sonnet 5.5 e413530658 Hold Play to move bank 2 up a layer in the SMC-PAD smoother.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-22 11:44:19 +13:00
jimmyandClaude Sonnet 5.5 6c060720ee Add two sets of colour scenes to the test show and light the pads from them.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-22 11:44:19 +13:00
jimmyandClaude Sonnet 5.5 b8e6b5f24b Add multipage frames and Playback sliders to the show builder.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-22 11:44:19 +13:00
jimmyandClaude Sonnet 5.5 82e75e69de Close MIDI ports properly in the pad editor.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-22 11:44:19 +13:00
jimmyandClaude Sonnet 5.5 3992744a1d Light pad bank 2 from page 7, and add factory reset and a pad bank toggle to the editor.
Bank 2's lamp colours come from preset page 7, not page 3, so pad_colours.py
writes there and the editor labels it Bank 2. The factory reset loads the
image MidiSuite uploads for its own reset, and the bank toggle writes the
global-block byte MidiSuite uses to switch banks.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-21 20:34:34 +13:00
jimmyandClaude Opus 5.5 63b3129cca Add an SMC-PAD preset editor that saves pad colours the way MidiSuite does.
The record layout and writes come from usbmon captures of MidiSuite:
a pad colour is one three-byte write to its preset address, which lights
the lamp straight away, and Save is the empty write that keeps it after
a power cycle. Pad records are channel, note, min and max velocity, RGB,
then Led. Tempo, swing, pad curve, and transpose come from the global
block, memory type 4. The factory image from MidiSuite's Reset is kept
as a test fixture, and scripts/decode_usbmon.py prints the SysEx in a
capture.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-21 19:30:34 +13:00
jimmyandCursor adca8e17a4 Put the XCHC colour scenes on the virtual console for the SMC-PAD.
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-10-21 17:39:13 +13:00
jimmyandCursor 24f9ca300e Add a stage simulator that draws the rig from Art-Net.
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-10-21 17:39:13 +13:00