ce933a4f321acdceb7d93838e7fddb3403ad33fc
344
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
  adminandClaude Opus 5.5
|
ce933a4f32 |
Let a drag change only the picked clip in the Sound and Video tab (#403)
A drag with a clip picked could span a new clip in the strip or span the picked one anew over the wave. Both happened by accident while working on a clip. Now a drag only makes a new clip while none is picked. With one picked it moves its edges or the whole of it. - With no clip picked, a drag in the strip or over the wave (in the Video tab the track) spans a new clip, as before. - A click over the wave or in the track picks the clip at that point, as a click on its band does. Beside the clips it picks the sound or the video. The picked clip wins where clips lie over each other. - With a clip picked, its handles move its edges as before. A drag on its band in the strip moves the whole clip, its length kept and inside the file. Any other drag does nothing. - The rules live in waveView.ts (clipGrip, clipToPick, movedSpan) and both tabs use them. The cursor shows grab over the picked band. - The Video tab's track draws the clips not picked faint, as the wave does, so a click there has something to aim at. Tests: - Unit wave-view.test.ts: 16 tests (4 new) - Counter-check: a strip drag beside the picked clip spanning a new one again fails the picked-clip rule test - e2e 34 + 66 (atx): ALL PASSED, 162 checks, 1 new check in each - Gate (atx): 295 files, 4527 tests Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
d48f174caf |
Play videos on sprites, converted on the server with clips (#402)
A video becomes an asset like a sound: uploaded once, converted to one format by a worker of its own, cut into clips in its tab, and worn by a sprite through its Texture row. Only Chrome is a target, without alpha. - Server: POST ?r=videos queues the upload under the hash of its bytes, so the MP4's name is known at once. The worker (VideoWorker over VideoFormat) writes H.264/AAC, yuv420p, constant frame rate, a keyframe every 15 frames, at most 1920x1080, with a poster and the details beside it. A failed job says why and is tried again on the next upload. - An upload may be up to 2 GB. It is streamed from the request to disk and hashed there, so php-fpm holds none of it in memory. The other routes keep their 256 MB. - The worker runs in the container game-maker-video on stein2, a compose project of its own (tools/video/), and in CI beside the PHP server. The CI image has ffmpeg for it. - Model: asset type video with size, length, frame rate and sound, and clips like a sound's. The clip helpers and commands serve both. - A sprite's Texture row offers videos and clips after the pictures. resolvePicture names a playing video by a key, and the texture library hands it to a VideoLibrary with one player per key. - The editor shows a placed video as a quiet preview. Play starts it over at its clip's start and lets it be heard through the master bus, a pause holds it, Esc makes it quiet again. - Video tab: the picture, a ruler, the clip strip and a track on the Sound tab's time axis. A drag spans a clip, a handle moves an edge, Play plays the picked clip. - The picker shows a video on its way with its progress and adopts an upload from it once it is ready. Dropping a video file does the same. Decided on my own: - Two sprites wearing the same clip share one player and show the same picture. Control per instance belongs to the follow-up with nodes. - A video plays in real time, the game's time scale does not slow it. - The container may use 3 of the 4 cores of stein2. Findings: - The first real upload (279 MB) got 413: the API's own ceiling of 256 MB for every request, not nginx or the Brandmauer. - Chrome has no picture of a video that never played, and a Pixi VideoSource drawn before the first frame stays empty for good. The texture is made in the first frame callback (openVideo). - A player nobody asks for in 2 s is let go. The inspector's PREVIEW asked only on an edit, so with the sprite not on screen its texture was destroyed under it and the batcher threw on every frame. The preview now redraws every frame while it shows a video, like a Time shader. A let-go texture keeps an empty source instead of none. - VideoSource.load() reloads the element and cuts off an earlier play(). Playing starts after loadeddata. Tests: - PHP videos.test.php: 68 checks, with the real ffmpeg - 300 MB upload through Brandmauer and nginx: 201 in 21 s, php-fpm at 13.7 MB RSS at most - Unit videos.test.ts: 20 tests - e2e 66 (stein2, atx was off): ALL PASSED, 18 checks; e2e 04 with the new asset count green - Gate (atx): 294 files, 4520 tests. serialize.test.ts used video as its example of an unknown type and now uses a made-up one. - Counter-checks: a keyframe gap of 30 fails the PHP check, a key that keeps the video's end fails 2 unit tests, a clip without its end fails "plays only its stretch" in e2e 66 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
6383cc0d8e |
Show a Sample node as a device that draws its waveform (#400)
The Sample was the last node of an instrument that still showed rows. It now draws the sound or clip it plays, so the loop and what plays after the note can be seen instead of read off two numbers. - SamplePanel, SampleFace and sampleView: a compact gold plate of 480x280 with the sound picker beside the name, a waveform screen, the four ways of playing it as buttons two by two, the root as a knob that reads as a note, a lever for following the note and the loop's start above its end. - The waveform comes from the library's peaks of the played sound, a clip's own cut included, scaled by the sound's volume. It is loaded on demand as the Reverb's sound is. - Each part has its colour: played once, looped, played after the note, never heard. The loop is held as playback holds it: an end of 0 or past the sound is the end, a start not before it loops all of it. - Idle settings come from the node's own idleParams, so the device greys out what the rows did. - A new instrument places its Output at x 840, past the device. At 460 it lay under the Sample. - Hook __gm.sample.panel and a check in e2e 34 behind the first instrument check. Tests: - Unit sample-view: 11 tests, new. pattern and chip-presets green. - e2e 34 (atx): ALL PASSED, 144 checks (1 new). - Gate (atx): 293 files, 4495 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
a6c00f7b8b |
Show a Bitcrusher node as a device that draws its pixels (#399)
A Bitcrusher only said what it does in rows of numbers. How coarse its steps and holds are could not be seen. Now the node is a device of its own size. Its screen sends a sine through the node's own crush and draws what comes out in cells, as tall as a step and as wide as a hold, so the coarser the crush, the coarser the picture. - CrusherPanel draws a 680x280 plate in bubblegum pink. Bits and rate are large knobs, wet is small. Rate gives its low end more of the turn, 8 kHz sits just past the middle. - crushTrace runs a 440 Hz sine at 0 dBFS through Crusher at 48 kHz and blends it as the worklet does. Where each hold starts comes from Crusher too: a ramp through it at 16 bits takes a new value per hold. - cellOf shapes a hold. From 5 px a step or hold gets a grid line and a gap. Finer steps give a 2 px bar. A cell is never more than twice as tall as wide, so 1 bit gives bars at -1, 0 and 1. - The screen shows two periods, or 6 holds where the rate is low. The sine going in is a grey ghost. Below wet 1 the heard line lies over the cells. - Above the screen a word for how crushed it is, CLEAN to BROKEN. Under it the noise the bits leave and where the rate folds back. - 16 lamps show the bits kept, the highest on the left. A CRUSH lamp beside the name glows the more it crushes. - COLUMNS_OF gives the Bitcrusher 4 columns in the presets. - Hook __gm.bitcrusher.panel(id) for the sound FX and instrument graphs. Decided myself: - The screen is still. Nothing on it moves, so tick() says false. - How crushed counts whichever of bits or rate takes more, the rate in octaves, times wet. Found: at 8 kHz the first hold is 7 samples long, the rest 6. Six sixths add up to just under 1 in Crusher. The worklet does the same, so the screen shows what is heard. Every sfxr effect preset ends in a Bitcrusher and gets 3 columns wider. The Laser preset still fits whole on Fit. Tests: - Unit: crusher-view 13 tests (new). The screen matches Crusher with its wet on three settings. - Counter-test: without the wet in the blend the worklet test fails. - e2e 34: ALL PASSED, 143 checks (1 new). - Gate: CHECK EXIT 0, 292 files, 4484 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
b6749cc43d |
Show Tremolo, Vibrato, Phaser and Flanger as devices that swing (#398)
The four effects that swing were still drawn as rows of numbers. Like
the Delay, the Saturation and the Limiter, each node now shows a device
of a fixed size. Its screen draws what it does over the first seconds
of a sound, and a light plays that in real time, over and over.
- One device for all four (ModulationPanel, ModulationFace and
modulationView), as KnobPanel is for Gain, Pitch, Mix and Pan. A short
description gives each its title, plate, buttons and knobs.
- Tremolo, 640x280 in cream: a wave its volume shapes, the volume as
its outline. It says DOWN TO 50%, EVERY 200 ms and DIPS -6 dB.
- Vibrato, 640x280 in plum: the pitch in semitones, cool below and warm
above. Under a wet below 1 the dry pitch stands beside it. It says
how late the voice is (LATE 0.93 ms).
- Phaser, 760x280 in mustard, and Flanger, 680x280 in steel blue: the
response from 20 Hz to 20 kHz, and under it a lane of the turn or the
delay over time. The response is drawn where the light is, so the
notches swirl and the comb jets. A comb finer than a column fills it.
- The light loops over 1, 2, 4 or 8 s, two LFO swings at least, for a
Flanger as long as its sweep takes to reach 20 ms. The face only
keeps the editor drawing while something swings and is heard.
- A Rate lamp beside each name goes from dark to lit with the swing.
- Nothing is listened to. The pictures use the worklets' own numbers:
lfoAt, vibratoSpeedUp (now exported from lfo.ts), the allpass chain
with its feedback, and the comb, reckoned at 48 kHz.
- Phaser and Flanger take 4 preset columns. Phased Pad was looked at
in a picture, nothing overlaps.
- Exported from nodes.ts: LFO_SHAPE, lfoShapeOf, TREMOLO_ROWS,
VIBRATO_ROWS, PHASER_STAGE_ROW and phaserStagesOf. FLANGER_MAX_MS is
exported from phaser.ts.
Decided here, as the request left the look open:
- A notch counts where the response falls below -3 dB, not some dB
under the loudest point. Feedback rings peaks far above 0 dB, and the
first rule counted 7 notches on a Phaser of 12 stages.
- Rate and Depth of Tremolo and Vibrato are big knobs, Wet is small.
Tests:
Unit modulation-view (stein2): 22 tests, new. Phaser and Flanger
give the dB of their worklets on a sine within 0.005 dB, the
Tremolo its samples, the Vibrato the pitch it reads at.
Counter-check: with the sign of the Phaser's feedback flipped in one
place, exactly the worklet test fails.
e2e 34 (atx): ALL PASSED, 142 checks, 4 new. Each device has its
size, its ports outside, a light that runs by itself, and a button
and a knob as one undo step each. Picking 8 stages on the Phaser
draws 4 notches.
Gate (atx): CHECK EXIT 0, 291 files, 4467 tests.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
adminandClaude Opus 5.5
|
28909f0adf |
Show a Limiter node as a device that holds what it took away (#397)
A Limiter only said what it does in rows of numbers. How much it took away, and when, could not be seen at all, and a short sound is over before any needle could show it. Now the node is a device of its own size, and its screen catches the reduction from the moment a sound starts and holds it once the sound ends. - LimiterPanel draws a 720x280 plate of brushed brick. The threshold knob is large, ceiling and release are small beside it, auto-gain is a lever. A LIMIT lamp lights while it takes something away. - The screen hangs the reduction down from the top, 0 to 24 dB by the square root like the Compressor's meter. The level coming in runs along the bottom, orange above the limit. Time runs from the start of the sound over 1, 2, 4 or 8 s, and a longer sound rolls. - ReductionCapture in limiterView holds the capture after 0.4 s of silence, without the silence, until the next sound. It writes the most taken away and when, and NOW or LASTED on the right. - The Limiter is tapped now, like the Compressor. The face reads the loudest sample since the last frame on both sides. - SpectrumTaps.levels takes a lateBy. The side in front is read 5 ms earlier, so a peak going in is met by what it became. - limiterReduction leaves the auto-gain out and solves Wet's blend for the limited part. Less than 0.2 dB counts as none. - A capture is also held when the node leaves the picture. Changing a knob or the lever leaves it as it is. - Hook __gm.limiter.panel(id) for the sound FX and instrument graphs. Decided myself: - The capture stays held until the next sound, not for a fixed time. - The screen has no wet knob, as the Compressor's has none. - The Limiter's two taps cost two gains of 1 per voice. The analysers only exist while the node is in the picture. Measured on stein2: the seeded blip at -4.4 dBFS against -12 dB showed 7.56 dB. Three hits at 0, -8 and -2 dBFS showed 11.9 dB at the first. In e2e 34 on melektaus-atx, -10.44 dBFS in against -20 dB showed 9.56 dB, the difference to the hundredth. Tests: - Unit: limiter-view 13 tests (new), spectrum 6 (1 new), soundfx 67 with the Limiter's taps. - Counter-test: without Wet's share in the formula the blend test fails. - e2e 34: ALL PASSED, 138 checks (1 new). - Gate: CHECK EXIT 0, 290 files, 4445 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
80b25488ce |
Show a Saturation node as a device that glows with its drive (#396)
A Saturation in the graph was a stack of rows. The user wanted a device of its own style and size, like the Delay and the Reverb, that shows warmth and heat the harder the sound is driven. - The node is a copper plate of 720x280 with a button per type, the auto-gain lever, a big drive knob and small knobs for tone, output and wet. Each change is one undo step, wired numbers lock their knob. - The screen draws a sine at -12 dBFS, the level auto-gain is reckoned at, as a grey ghost. Over it is the wave that comes out: bent by the type, its DC taken out, auto-gain and output after it, blended by wet. Under it stand the auto-gain and the peak. - The heat is the drive over 48 dB, as far as wet lets it be heard. The wave goes from cold steel through ember red to white, a glow rises from below and flickers, and embers fly up from about 8 dB on. A lamp beside the name glows with it. - The node keeps the editor drawing only while there is heat. Without drive it draws nothing, unlike the Delay, which always animates. - The Reverb's private type button is now PickButton with its own colours, and the Saturation's type buttons use it as well. - No instrument preset uses a Saturation, so no preset needs more columns. Tests: - Unit saturation-view: 14 new tests. - Counter-check: without taking out the DC, exactly the DC test fails. - e2e 34 (atx): ALL PASSED, 137 checks, one new for type, lever and drive with undo. - Gate (atx): CHECK EXIT 0, 289 files, 4431 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
d4a45c127b |
Show a Delay node as a device with its echoes over time (#393)
After the Chip, the Envelope, the Reverb and the one-knob devices, the Delay is drawn as a device of its own size too. Its screen shows what time, feedback, damping, wet and ping-pong make of a sound, and a light runs over the echoes at the beat they come at. - DelayPanel: a 680x280 plate in the Compressor's brushed aluminium, now exported as ALUMINIUM. A screen, levers for Sync and Ping-pong, knobs for Time, Feedback, Damping and Wet, and an Echo lamp. - The screen draws the sound as a white bar and each echo as a bar one step later. A bar is as tall as the echo is loud, evenly in dB to -60. It turns from warm white to rust as the damping takes the highs. With ping-pong the bars alternate above and below, the first left. - At most 24 bars are drawn, and the screen says how many more are heard. Under it stand the step and the end of the tail, above it where the highs of the first echo end. - The running light: DelayFace.tick returns true while the node is in the picture, so the editor keeps drawing, as for the Particles tab. The light goes round the span in real time, a bar glows as it is passed, and the lamp blinks with every echo but not with the sound. - On beats the screen counts at 120 BPM, as the tempo of the voice is not known in the graph. The time knob then names the note, dotted and triplet ones included. In seconds, 2 s and up reads "2 s max". - delayView holds the rules apart from the drawing: layout, tapers, echoTrain, echoCutoff for k rounds of the Butterworth low-pass, sweepAt and the texts. At wet 0 the rest is drawn faint. - delayEchoes in the core counts the echoes heard. The tail of a voice and the screen use it both. DELAY_ROWS, DELAY_SYNC and PING_PONG are exported. - Knob.reformat writes the number anew when its format changes, here when Sync switches between seconds and beats. - The test hook __gm.delay.panel returns knobs, levers, idle settings, the echoes drawn, the texts and where the light is. - Instrument presets give the Delay 4 columns. Echo Lead and Bell use one. In the picture of Echo Lead nothing overlaps. Found on the way: a node with a face draws no setting rows. The e2e check of a bypassed Delay (#354) read "Bypass: true" from them. It now checks the faded node and the stages heard, as before. Tests: - Unit delay-view: 17 tests, new. Counter-check: the first ping-pong echo on the right turns the side test red. - e2e 34 (melektaus-atx): 136 checks, 1 new. - Gate (melektaus-atx): 288 files, 4416 tests. - Pictures looked at: default, ping-pong, beats with damping, 83 echoes at 50 ms, wet 0, and the preset Echo Lead. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
389d9555fa |
Show a Pan node as a one-knob device in cyan (#392)
After the Gain, the Pitch and the Mix, the Pan is the next node that is just one number. It now gets the same device in the graph instead of a row for its pan. - PAN_DEVICE: the Gain's 160x160 plate in cyan with one bipolar knob for Pan. The middle stands straight up, and the turn goes evenly to L and R, which are engraved at its ends like A and B on the Mix. - PAN_TAPER in hardwareView is 1, so left and right are heard alike. Device and unit test read the same constant. - PAN_ROW is exported from the core, so the knob has the node's range. - A linked Pan greys the knob, as for the Gain. - The test hook __gm.pan.panel returns the same shape as the Gain's. - No instrument preset uses a Pan, so no preset columns change. Tests: - Unit hardware-view: 17 tests, 2 new (the turn, and L and R inside the plate and clear of the screws). Counter-check: PAN_TAPER 2 turns the turn test red. - e2e 34 (melektaus-atx): 135 checks, 1 new. - Gate (melektaus-atx): 287 files, 4399 tests. - Pictures looked at: Gain, Pitch, Mix and Pan in a row, Pan at -0.35. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
fe4e0d9f1a |
Snap to the grid from the start in every graph tab (#391)
Laying out a graph by hand mostly means lining nodes up, so the magnet was switched on in nearly every session. Every graph tab now starts with it lit: Behavior, State machine, Shader, Particles, Sound FX, Instrument, the song's Routing and the Routing tab. - SnapToggle takes the state it starts in and lights its button to match. GRAPH_SNAPS_AT_FIRST sets it for the toggles and for GraphCanvas alike, so button and canvas cannot disagree. - Curve and Vector Art are no graphs and still start without it. - New hook __gm.graphSnapping(on?) reads every graph tab's magnet, or "mixed" where button and canvas differ, and switches them all. - The e2e harness reads the magnets after each navigate() and then switches them off, as it already sets the zoom transition to 0. Most checks drag a node to an exact spot. e2e 45 does the same itself, because the editor comes up there through the sign-in. - e2e 06 checks that all eight graph tabs started with the magnet on. The checks in 06, 26 and 65 that read it off now say they read what the harness left. Tests: - Unit snap-toggle: 4 tests, 1 new (lit at first, nothing asked of the target before a click). - e2e 06 (melektaus-atx): 150 checks, 1 new. e2e 34: 134 checks. - Gate (melektaus-atx, together with #390): 287 files, 4397 tests. - Pictures looked at: none, the state comes from the hook. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
e70954115c |
Show Pitch and Mix nodes as one-knob devices in green and blue (#390)
The Gain has been a small device with one big knob since #384. The Pitch and the Mix are just as much a single number, so they now look and work the same way. The Mix also gets a lever for its curve, which until now was only a row in the inspector. - GainPanel and GainFace become KnobPanel and KnobFace. Each node is described by a KnobDevice: its plate, its row, how the knob turns, and an optional lever. The Gain looks as before. - Pitch: a green 160x160 plate with a bipolar knob for Semitones. 0 stands straight up and taper 2 puts an octave at 43 % of each half, so small steps near 0 are easy and the full +-64 is still there. - Mix: a blue 180x200 plate, the lever on top and the knob below. The knob has A and B engraved at the ends of its turn. The lever reads "crossfade" down and "both at center" up with its lamp lit. A flip is one undo step. - A wired Semitones or Mix greys the knob, as for the Gain. - The instrument presets give Pitch and Mix two columns. In the graph they are 281 and 261 units wide, more than one column of 240. - The test hooks __gm.pitch.panel and __gm.mix.panel return the same shape as __gm.gain.panel, plus the lever. Tests: - Unit hardware-view: 15 tests, 3 new (Pitch turn, lever above the knob and ends on the Mix plate). Counter-check: taper 1 turns the Pitch test red. - e2e 34 (melektaus-atx): 134 checks, 1 new. - Gate (melektaus-atx, together with #391): 287 files, 4397 tests. - Pictures looked at: Gain, Pitch and Mix in a row, Pitch at -24.5, Mix with both curves, the Bell preset without overlaps. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
e8021b1fd3 |
Show a Reverb node as a device with a round screen of its tail (#389)
The Compressor, the Gain, the Chip and the Envelope are devices in the graph. The Reverb was still a list of rows, and how long and how dark its tail rings was left to the numbers. - The node is a plate of 640x280 in pale lavender. Its inputs are on the left, its output on the right. - A round screen draws the tail as rings out from the sound. A ring's radius is its time, as the root of its seconds, on one scale up to 10.5 s for every Reverb. So a Room is a small disc and a Cave fills two thirds. - A ring is as bright as the tail is loud there, evenly in dB. Its colour is how much of the highs is left, from ice to violet, with the cut-off of the impulse response (reverbCutoff, now shared). - The pre-delay is a dark gap round the dot, a dashed ring marks -60 dB. Seeded dots stand for the reflections. Wet fades the tail against the dot. - A type of Sound draws the peaks of its sound from the library, each against its loudest. It says LOADING until they are there. - A button for each type, a list of the sounds and four chrome knobs: Decay, Damping, Pre-delay and Wet. - A knob the type does not take is greyed but turns. It goes over to Custom with the type's numbers, as one undo step. Picking a sound makes the type Sound, also one step. - NodeParamDrag can write several settings as one drag for that. - Test hook __gm.reverb.panel(nodeId), for the sound FX and the instrument graph. Tests: - Unit reverb-view: 10 new tests for the plate, the knob turn, the numbers per type, the scale, the tail and the texts. A screen that ignores damping turns the highs test red. - e2e 34 (atx): 1 new check (device between its ports, decay knob of a Hall goes to Custom as one undo step, Cave button, a sound from the list makes it Sound and draws that sound). - Probe pictures on stein2: Room, Hall, Cave, Plate, Custom with damping 1 and pre-delay 0.3 s, Sound. - Gate (atx): CHECK EXIT 0, 287 files, 4393 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
c75e72d4ee |
Show an Envelope node as a device with its ADSR curve (#388)
The Compressor, the Gain and the Chip are devices in the graph. The Envelope was still a list of rows, and its shape was left to the imagination of whoever read the numbers. - The node is a plate of 520x280 in pale sea green. Its inputs are on the left, out and value on the right. - A dark screen draws the curve over a note in amber. A letter sits above each stage, its time below it, and the sustain in percent above its line. - The time on the screen is squeezed. Each stage is as wide as the root of its seconds, at least 3 %, and the hold has a fixed fifth. So 5 ms of attack still shows beside a release of 3 s. - The falls use SETTLE, as the player does. With sustain 0 there is no hold and no release, and the screen says NOTE ENDS. - Attack, Decay, Sustain and Release are chrome knobs. The times use taper 4, so 5 ms sits an eighth of the way and 1 s not quite half. - From and To have no range a knob could turn through. They are NumberFields on the plate, typed in or scrubbed. - The decay at sustain 1 and From/To while nothing takes value are faded. While a link takes value, the bezel says what it moves between, as VALUE 400 TO 3500. The UI font has no arrow. - Instrument presets give an Envelope 3 columns of 240. - Test hook __gm.envelope.panel(nodeId). Tests: - Unit envelope-view: 6 new tests for the plate, the knob turn, the idle settings and the curve. A linear decay turns two of them red. - e2e 34 (atx): 1 new check on the Laser preset (device between its ports, curve ends with the decay, decay knob and scrubbed To field each one undo step). - Probe pictures on stein2: Synth Brass preset, pad, pluck, organ and an Envelope whose value moves a Chip. - Gate (atx): CHECK EXIT 0, 286 files, 4383 tests. The first run was red on the font test (#214) for an arrow in the bezel text. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
eb78f3cbf3 |
Show a Chip node as a device with its wave and its pitch (#386)
The Chip had eleven numbers in rows and nothing that showed what they do to a note. Its node is now a device of one size like the Compressor's, and two screens draw the wave and the pitch from the settings, worked right in the instrument graph. - ChipPanel is 840x280 in the warm grey of a handheld console. A button per wave, the "follow note" lever and a chrome knob per number, grouped Oscillator, Slide, Vibrato, Arpeggio and Repeat. - The wave screen draws two periods as loud as the Chip plays them, a square at its duty. toneValue and periodicSteps come out of chip.ts, so the screen and the worklet share one formula. - The pitch screen draws what slide, delta slide, vibrato, arpeggio and repeat make of the pitch over a note. It ends where the note falls below "min frequency" and starts from A4 while the Chip follows the note. - Idle settings are faded, wired ones greyed and locked, as the rows were. Every change is one undo step. - Knobs take a bipolar range: 0 straight up, the taper bends each half. Slide, delta slide, duty sweep and arpeggio use it. - The Compressor's auto-gain lever is now a Lever with a name of its own, which the Chip uses for "follow note". - Presets give a Chip five columns, as a Filter has four since #385. Found on the way: - With the Chip this wide the Laser preset is about 3060 units. In a 1280 px window Fit stops at the least zoom of 0.25 and cuts the Notes node off. The zoom limit is left as it is. Tests: - Unit: chip-view (9 new), knob (1 new), chip, chip-presets and hardware-view, 53 green - e2e 34 on melektaus-atx: ALL PASSED, 131 checks (1 new). The first run was red on the check itself: a Laser's "min frequency" can lie above A4, and the screen then rightly says the note ends at once. - Gate on melektaus-atx: CHECK EXIT 0, 285 files, 4377 tests - Counter-check: a square drawn with its duty mirrored turns the wave test red Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
c087993000 |
Edit a Filter in its node and resize EQ and Filter nodes (#385)
A Filter sounds like one band of the Parametric EQ, but it keeps what the EQ cannot: linked numbers, the all-pass and +-40 dB. So it stays a node of its own and gets the EQ's editor with itself as the one band. The editors also stop running out of their room: what does not fit is left out instead of being cut. - The curve of the EQ is now BandCurve with a host the EQ and the Filter both fill. CurvePanel holds what their panels share. - FilterPanel shows the Filter as one band. A drag of its handle sets frequency and gain as one undo step. A number fed by a link is greyed out and the drag leaves it alone. - A double-click on the Filter's title opens the same in a tab of its own. Delete there does nothing, the node stays. - The Filter is tapped like the EQ, so its node and its tab draw the spectrum of what passes it. A tap is two gain nodes per voice, the analysers are made only while someone listens. - EqFace and FilterFace are one CurveFace now, given the panel. - EQ and Filter nodes have a corner to resize them, drawn like the floating window's. The size is saved on the node as `size`, each drag is one undo step, the least is 240 x 120. - In the node the help line is gone. The row of numbers comes in only once the curve keeps 120 px of height. A group of the row that does not fit is not drawn, the one on the right first. - A frequency or dB label that would run into the one before it is left out. - The all-pass is drawn flat. The curve's maths takes every shape a Filter has. Found on the way: - The instrument presets set nodes 240 apart. A Filter with its editor is about 700 wide and covered the Bitcrusher, so a Filter now takes four columns. Tests: - Unit: 8 new (filter band, resize command and save, fitting rules), stage lists with a Filter now hold its taps - e2e 34 (atx): ALL PASSED, 130 checks (4 new) - Gate (atx): CHECK EXIT 0, 284 files, 4367 tests Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
01ee61868b |
Show a Gain node as an orange device with one knob (#384)
A Gain node now draws a small device in the graph, as the Compressor does since #380: the brushed plate in orange, 160x220 graph units, with one chrome knob between its inputs and its output. - GainPanel and GainFace join the node faces. The knob drags, resets on a double-click and keeps the wheel for the graph (#383). Each turn is one undo step, a linked Gain (dB) greys it. - The knob turns with a taper of 0.26, so 0 dB stands straight up and the range below -24 dB takes a fifth of the turn. - drawPlate, engrave and the brushed look move out of CompressorPanel so both devices share them. The Knob takes an arc colour, because the orange arc vanished on the orange plate. - No meter and no tap. Gain sits in every voice of an instrument, and a tap would add two audio nodes to each note for a display nobody asked for. Found on the way: - Knob.dialGlobal added half the dial unscaled. In a zoomed graph it pointed beside the dial. The Compressor's e2e only hit by luck. Tests: - Unit: 2 new for the Gain plate and its taper. - e2e 34 on atx: ALL PASSED, 126 checks (1 new). - Gate on atx: CHECK EXIT 0, 283 files, 4359 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
79a04e6764 |
Edit a Parametric EQ in its node, let the wheel only zoom graphs (#381)
The small curve below the EQ node's rows (#375) only showed the bands. Like the Compressor since #380, the node now shows the editor of its tab between its inputs and its output, and is worked right there. A double-click on the node still opens the same editor in a tab. Two wishes for the nodes with a panel came on top and share the same files, so they are in this commit as well: the numbers below the Compressor's knobs read as plain labels (#382), and the wheel over a node zooms the graph instead of turning a value (#383). #381, the EQ in its node: - EqPanel holds the curve, the band's numbers and the help line, taken out of EqEditor. The tab puts its one panel under a header, each EQ node has a panel of its own at 720 x 360 (EqFace). - A press on the curve neither picks nor moves the node. A double-click on the curve adds a band, one on the title or the edge opens the tab. - In the node a press on the Type or Slope button stays with the button, so the node is not picked or moved. - The spectrum runs in the node while it is in the picture. The tab lets go of the tap once, like the Compressor's, and Editor.tick lets every unseen tab and graph go before any seen one listens. - The count of bands moved from the tab's header into the band row, so the node shows it as well. - The node picture of #375 is gone with its only user: NodePicture, EQ_PICTURE and NodeView's drawing of it. #382, the numbers below the Compressor's knobs in its node: - A Knob with typed: false shows its number as a centred Label without a box, in the panel's ink. The ratio reads 4:1 (ratioText). - The label is as high as its text and sits up in the dial's open bottom, right under the ends of the scale. - Label takes a colour of its own. #383, the wheel in a graph: - GraphCanvas takes the wheel on the way down and stops it there. It zooms over a node too, and no knob or curve in a node sees it. - The tabs keep their wheel: a knob's step, a band's Q and the zoom of the EQ's dB. Found on the way: - A Menu opened below a button in the zoomed graph would have been placed by the button's own height. openBelow now takes the bottom edge as drawn. - The band row did not lay out its labels after a text change when its own size stayed. In the tab a later resize hid it, in the node "Band 1" and the count were missing. The row now lays out itself. - Handle and point positions for the checks were unscaled. They now go through toGlobal and hold at the graph's zoom. Tests: - Unit parametric-eq: 16 tests. The #375 tests now check eqCurvePoints and eqRangeOf directly. compressor-view: ratioText. - e2e 34 (atx): ALL PASSED, 125 checks (3 new: the EQ panel and its ports, editing in the node, the spectrum in the node with the tab open). The Compressor node's check reads its numbers as labels and 8:1. Both node checks now expect the wheel to zoom the graph and leave the ratio, the Q and the dB zoom alone. - The unseen-tab check now leaves for the Scene tab, since the node in the graph listens to the same tap. - Counter-check: the tab letting go of the tap every unseen frame stalls the node's spectrum with the tab open (1 frame against 22, probe on stein2). - Gate (atx): CHECK EXIT 0, 283 files, 4357 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
9917253751 |
Open a Compressor as a device in a window of one size (#379)
The tab from #378 shows a Compressor as an editor. The user wanted it to look like a device as well, after a brushed aluminium plug-in, with a meter of the gain reduction instead of the pictures. The tab stays and is switched to per node. - Hardware look is a switch under Settings, on by default for new and existing Compressors. Only the editor reads it. - The panel is 820x280 px and opens straight in a window of that size. The window has no sizing corner and no Dock. A pane says so through `fixedSize`, and `keptSize` keeps the window's middle when it changes to that size. - A Compressor in this look never sits in the tab row. Opening one floats it, and switching the look on moves its tab into the window. - The panel has chrome knobs (threshold and makeup large), a meter with a needle that moves like a VU's, an LCD with the name and the reading, a lever with a lamp for auto-gain, and LEDs for the level going out. Everything is drawn in code. - The Knob takes a size and a chrome look, and leaves out its caption when it has no name, so the panel engraves its own. - `autoGainOn` is now `switchOn`, because the new switch reads the same way. - Where parts sit, the meter's scale, the needle's speed, the LEDs and the plate's grain are in `hardwareView.ts`, apart from the drawing. Decided myself: - The meter's scale runs by the square root, so the first dB is about as wide as the last nine. It is red from 10 dB on. - The LEDs light one every 3 dB from -39 to 0 dB and fall by 20 dB a second. - Switching the look changes the graph's fingerprint, like any other setting of a node. The sound FX is built again once. Findings, fixed on the way: - Pixi's `arc` goes on from the last point of the path. The meter's scale had a line from the corner of the face until each arc got a `moveTo` to its own start. - A text put into a Graphics makes Pixi warn that only Containers take children. The LCD is a Container now. - The first browser check read the needle at one moment and missed the short sound. The panel keeps the most its needle and LEDs showed. Tests: - Unit: hardware-view (10 new), documents (1 new) - e2e 34 (melektaus-atx): ALL PASSED, 122 checks (4 new). Played, the needle reached 20.9 dB and 5 LEDs lit. - Gate (melektaus-atx): CHECK EXIT 0, 283 files, 4355 tests (11 new) - Counter-checks: a window that keeps its corner instead of its middle, and a linear meter scale, each turn exactly their test red. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
2d4909d248 |
Fit the Compressor's tab to any size, make auto-gain a switch (#378)
Two things came up on the new tab. Floated and sized down, it spilled its knobs out of its window, and sized up it left the window empty. Auto-gain lifted far too much: it brought 0 dB back to 0 dB, which at -40 dB, 4:1 and knee 6 is +28.1 dB, so a sound at -24 dB came out at -6 dB. And as a knob from 0 to 1 it blended between makeup and auto-gain, so at 1 the makeup knob did nothing. Layout: - The knobs stay on the bottom line, centred, with the help line above them. - The levels fill the height left above. The curve is a square at their right, centred up and down. It grows with the tab from 180 to 360 px and takes no more than 40 % of its width (picturesLayout). - A floated window cannot be sized below what its pane needs. A pane says so in its minWidth and minHeight, and the Compressor's is every knob and the smallest pictures. Other panes keep 320 x 200. - A pane put into an open window that is too small grows the window. Auto-gain, as decided in #378: - A Compressor's auto-gain gives back what it takes from a sound at -18 dB, so that one comes out as loud as it went in. At the numbers above that is +14.6 dB. - Auto-gain is a switch on the Compressor, the Limiter and the Saturation alike, a setting and no longer an input. The makeup or output is added on top of it instead of being blended with it. - A project from before reads a number as on above 0. Nodes never set keep their defaults: off on the Compressor, on on the other two. - The tab shows the switch at the right of the knobs, with the dB it gives below it. With it on, the curve has a line at -18 dB from the curve up to the diagonal. - Every level reading keeps the gain of its time. Before, the whole picture was read by the current makeup, so after a change the tinting of the last seconds was wrong or missing. Findings: - My first e2e run of the layout clicked Dock where it was before the window was sized up, so the window stayed open and the next check went red. The check reads the button again. Tests: - Gate (melektaus-atx): 282 files, 4344 tests (6 new) - e2e 34 (melektaus-atx): ALL PASSED, 118 checks (2 new) - Offline in Chrome: a sine at -18 dB through -24 dB and 4:1 came out at -17.92 dB with auto-gain on and -22.42 dB with it off. - Counter-checks: the curve set at the top instead of the middle, and every reading read at a gain of 0, each fail exactly their test. - Looked at the window in screenshots: at its least every knob and the switch show. At 1280 x 797 the levels fill the height, the curve is centred at the right and the knobs are at the bottom. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
1df38bd430 |
Edit a Compressor in a tab of its own, with knobs and levels (#378)
A Compressor had only its number rows in the node. Its numbers are
easier to set by feel on a dial, and what it does is easier to judge
when the level going in and the level going out are seen side by side.
- Knob is a new widget in the toolkit: a slider bent into a circle,
its name above and a number field below. A drag up or down turns
it, Shift a tenth as far, a notch of the wheel a fiftieth of the
turn, a double-click sets the default, the number takes a typed
value. A taper gives the low end more of the turn.
- The Compressor's tab opens on a double-click, one per node, like
the Parametric EQ's. It has a knob for each of the seven numbers,
greyed while its input is linked.
- Its curve shows level in to level out from -60 to 0 dB. The point
where it bends sets the threshold, the end of the curve sets the
ratio that ends there.
- Its levels show the last 4 s on both sides of the node. Where it
takes more than 0.5 dB away, the stretch below the level coming in
is tinted. The makeup is left out of that.
- The Compressor is tapped like the EQ. SpectrumTaps listens to more
than one node at once and reads the loudest sample of the last
25 ms on each side.
- The node tabs share one map in the editor (nodeTabs), so a third
kind needs no new plumbing.
Findings:
- CI task 614 for
|
||
|
adminandClaude Opus 5.5
|
23922ee4a2 |
Show the zoom of the music tabs and the EQ as a transition (#377)
The zoom transition of #309 left out the Pattern, Song and Sound tabs
and the Parametric EQ's curve. There a wheel step or Fit still jumped,
whatever "Zoom (ms)" in the general settings said.
- axisTween wraps ViewTween for a single axis. The music tabs zoom
their time across and their rows, tracks or height down apart, so
each keeps one tween per axis and either goes its own way.
- Fit by its button is shown as a transition too, as in the other
tabs. The fit on opening a document still jumps.
- A middle-button pan first lands a transition under way, and saving
a document's view lands it on the step it was heading for.
- The EQ's range is saved on the node at once. While the zoom is on
its way the curve draws a range of its own and writes nothing.
Its dB lines are those of the new range, sliding to their place,
so no line shows a crooked number.
Found on the way, not part of #377:
- CI task 611 for
|
||
|
adminandClaude Opus 5.5
|
82e96e680f |
Decode a clip from a file of its own, not its whole sound (#349)
A clip was cut from its sound's file, so a game decoded the whole file for it and held it for as long as the page was open. A long recording played through a few short clips cost the recording. Every clip now has a copy of its cut as Opus, with Stream on or off, and a game decodes that instead. - StreamKeeper makes the copy of each clip whether its sound streams or not. Stream only decides whether a sound gets a copy of its own. - clipCopyTable says which cut decodes from which copy. The sound library loads the copy for a clip unless the whole file is decoded already. Then the clip is cut from that as before. - projectFiles decodes the whole file ahead only where something may play the sound itself: no clips, a clip without a fitting copy, or the sound's id anywhere in an asset that can name one. That last test walks the data, not the fields, so no place is missed. - A copy that cannot be had falls back to the whole file. A library that bakes keeps to the whole files, as the bake's fingerprint does. Tests: - Gate: CHECK EXIT 0, 280 files, 4319 tests (7 new: streams +4, sound-library +3). - e2e 34: ALL PASSED, 108 checks (2 new). With Stream off a fresh library decodes the ten-second reversed clip's copy and not the five-minute file. It is the cut made from the original to within 0.0091 a sample and 0.013 dB in any 20 ms. Once a node plays the sound itself, the whole file is decoded ahead again. - e2e 65: all passed with this change. - Counter-checks red as they should be: the library ignoring copies, the whole file always decoded ahead, clip copies only with Stream. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
2764919bb8 |
Stream long sounds from an Opus copy instead of decoding them (#349)
A sound was decoded whole as soon as the project opened, in the editor and in a released game: five minutes of stereo held some 115 MB for as long as the page was open. A sound with Stream on now plays from a copy as Opus in Ogg, fetched a piece at a time and decoded a few seconds ahead by WebCodecs, so a game holds a few megabytes for it. - The Stream switch is a row of the sound, off by default. The Sound tab suggests it from a minute on and says how far the copy has got. - StreamKeeper makes the copies in the browser, a moment after the last change, one at a time: the sound's own and one per clip of its cut, fades and Reverse baked in. A copy that is no longer wanted is taken off and deleted once the project is saved without it. - streamTable says which played names stream from which copy. A clip without a fitting copy, a Sample and a Reverb impulse stay decoded. A sound with Stream on is never decoded ahead. - The page fetches by Range and takes the whole file from a server that ignores it, as the CI's PHP server does. It keeps the bytes and the first two seconds decoded, the rest goes four seconds ahead. - A jump decodes 32 packets before the place first. Measured in Chrome, what follows is then the same to the bit as decoding from the start. With 4 packets it was 0.06 off. - The stream-player on the FX worklet reads the samples at the rate of the context along a Catmull-Rom curve. What comes late is silent and skipped once it is there, so a clip keeps its beat. - The opening stretch goes with the player as it is made. Sent by message it came a few milliseconds late, and 557 samples of the attack were lost. - Pitch on a streamed sound is a shifter after the player, as on a bus: the length stays and it is heard 42.7 ms late. - Bakes of patterns and songs (#340) always stream now. - The song timeline draws a streamed sound from its copy decoded at 8 kHz, keeping only the peaks. Tests: - Gate: CHECK EXIT 0, 280 files, 4312 tests (31 new: ogg +3, stream-queue 6, streams 11, streaming 6, sound-library +5). - e2e 34: ALL PASSED, 106 checks (7 new). Five minutes of mono, 4.8 MB as WAV, 3.6 MB as its copy: streamed the renderer grows by 5.8 MB, decoded by 69.6 MB. Offline at 48 kHz, start, loop seam and a reversed clip are the copy decoded whole to the sample (0). - Counter-checks red as they should be: no pre-roll, no catch-up of late samples at the end, no opening stretch, a streamed sound decoded ahead. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
84cd36f902 |
Bake a pattern or song to Opus, played in its place by a game (#340)
A pattern or song played by a game runs every note through the scheduler, the instruments and their FX, live. Bake works that out once, keeps it as a file on the asset, and Play pattern and Play song play the file instead, as one voice, for as long as nothing it is made of has changed. The tabs always play live. - Bake is an icon in the header of the Pattern and the Song tab, lit while the bake holds. A click bakes, cancels a bake that runs, or offers Bake again and Remove bake. The toolbar shows the progress. The info of the tab says "baked" or "bake outdated". - The bake holds everything up to the output of the pattern or song. The sound FX of the node, fx params, volume, fade in, bus, Master routing and stopping with a fade stay live. - A print of the worked-out plan decides whether it holds: notes and clips, tempo, length, loop, tracks, the build of each instrument for each note, sound FX, files, cuts and levels. A rename keeps it, an undo back to the baked state makes it hold again. - The file holds the straight run and then one more round of the loop as heard the second time, tails included. Looped, it plays to the loop end and then round that round, to the sample. Once, it fades out over its last 10 ms. Let ring fades it out over 0.3 s. - It is worked out by a second SoundLibrary on an OfflineAudioContext, its scheduler woken by suspends of the offline clock, and what reaches the master is taken. A worklet probe tells whether that was mono, so a mono bake goes on exactly as the notes did. - Opus in Ogg, 160 kbit/s stereo or 96 kbit/s mono, encoded with WebCodecs, the Ogg pages written by hand. Normalized to full scale, the peak kept and played back. - The file goes up into the sounds of the account as "[bake] name", which the sound picker leaves out. An old bake file is deleted once the project is saved without it. Only a bake that holds is loaded. - Ogg files are decoded at 48 kHz: resampled to 44.1 kHz, the loop of a bake lost a sample a round. - Baking and removing a bake are no undo steps, like the EQ's zoom. Findings: - Chrome's AudioEncoder refuses opus.format 'ogg' (NotSupportedError) but hands out the OpusHead with a pre-skip of 312 for 'opus'. - The Song tab's header is full at 1280 px: a captioned Bake pushed Fit off the tab. Hence the icon. Tests: - Unit: tests/core/ogg.test.ts (6), tests/core/bake.test.ts (16), four new in sound-library.test.ts, one in song-play.test.ts. One test cuts, reverses and fades a clip on a song's track: each change outdates the bake, each undo makes it hold again. - e2e 64 (melektaus-atx): ALL PASSED, 36 checks (6 new). Live and baked rendered offline differ by at most 0.06 dB over 114 windows of 20 ms, two loop ends among them. e2e 65 had no red check with the icon in its header. - Gate (melektaus-atx): CHECK EXIT 0, 277 files, 4281 tests. - Counter-checks: print without instrument builds, print taking the sound's file instead of the clip's cut, round taken from the first pass, bake voice without its looping source, Play pattern or Play song ignoring the bake: one unit test red each. Pre-skip set to 0: the e2e comparison 4.8 dB off, only that check red. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
3377524958 |
Zoom the Parametric EQ's dB with the wheel, saved on its node (#376)
Close moves of a few dB were hard to set and to see on a curve that always spans -24 to +24 dB. The wheel on free space now zooms the dB the curve shows, and the node's small curve shows the same range. - The range is one of ±3, ±6, ±12, ±24 and ±48 dB, always the same either side of 0. Up zooms in. - It is saved on the node as `range`, so it is back after loading. Like the game scale (#226) it is no undo step. EditorState gets `setNodeView` for such a setting of a node. - The lines across the curve follow the range: every 1, 2, 3, 6 or 12 dB. A handle whose gain lies beyond the range stays on the edge, in sight and in reach. - The wheel over a handle still turns the band's Q. It no longer does so anywhere while a band is picked: after opening, a band nearly always is, and the zoom would be out of reach. - `EqArea` carries the range, so the tab and the node map dB alike. The spectrum keeps its own scale. Tests: - Unit parametric-eq: 16 green, 5 new (range in compile and save, node picture, wheel steps, grid lines, handle on the edge) - Counter-check: the node's picture without the saved range turns the new picture test red - e2e 34 (melektaus-atx): ALL PASSED, 99 checks, 1 new - Gate (melektaus-atx): CHECK EXIT 0, 275 files, 4252 tests Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
687377682c |
Draw the EQ's curve in its node and a spectrum behind it (#375, #374)
Both were asked for in #367 "for later". The node of a Parametric EQ now shows its curve in small, and its tab shows what passes the node while it is seen. #375, the curve in the node: - NodeDef has an optional `picture`: a height, and shapes made from the settings as plain point lists (guide, area, line). It is general, the Parametric EQ is its first user. - NodeView draws it in a box below the rows, again only when the box or the settings change. A drag in the EQ tab changes them, so the node follows it live. - Without a band the box stays, flat on its 0 dB line, so the node does not grow at the first band. - The axes of the curve, `eqCurvePoints` and `fillShapes` move from editor/soundfx/eqView to core/soundfx/parametricEq. Tab and node draw the sum from the same function. #374, the spectrum: - A node marked `tapped` gets a `tap` stage in front of its own stages and one behind them when compiled, in a sound FX and in an instrument alike. Its id is asset and node, since a copied asset keeps its node ids. - A tap is a gain of 1. SpectrumTaps (new, viewport/spectrum.ts) holds the tap of every voice and bus while it plays. - While the EQ tab is seen, in the tab row or in a window, the taps of that node feed one analyser per side (8192 points). Nothing is made before a tap of the node plays, so no AudioContext is opened early. - The tab draws both behind the curve, on a scale of 0 to -96 dBFS: the sound coming in faint and grey, the sound going on in the colour of a sound port. The analyser smooths it, and once the sound has gone the sum still hands it silence, so it falls away. - Once the tab is not seen, Editor.tick lets the analysers go, like the meters of the Song tab. - A tap does nothing to the sound, so `stats().last.stages` leaves it out, and a bypassed EQ keeps its taps. Decided on my own: - Every sound passing the node counts: Play of the sound FX tab, an instrument's keys, a song, a run. The issue proposed that. - The spectrum has no labels of its own. Its scale is in the wiki. Tests: Gate (melektaus-atx) 275 files, 4247 tests (11 new), EXIT 0 e2e 34 (melektaus-atx) ALL PASSED, 98 checks (2 new), twice Counter-checks, each put back by copy and compared with cmp: - The node's curve drawn without its bands: 2 unit tests fail. - The spectrum without the loudest bin between two points: 1 fails. - SpectrumTaps.stop leaving the taps linked: 1 fails. - Taps that never feed the analyser: e2e 34 fails in exactly the spectrum check (frames 0). Run in the container only. The screenshots of the second run were looked at: the blip as one peak behind the curve, and the dip at 4 kHz in the node. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
25084f8576 |
Refactoring round 10: split SoundLibrary, fold sound duplicates (#85)
Refactoring round 10 covers everything since round 9 ( |
||
|
adminandClaude Opus 5.5
|
3f12a77aea |
Preview a sprite at the end of its inspector, placed or as an asset
A sprite's inspector listed its rows but never showed what they add up to. The picture, its frame, nine-slice, tint and shader could only be judged in the scene, often far from the object being edited. - SpritePreview draws the sprite through syncSpriteMesh, the function the scene draws it with. Slices, padding, tint and shader come out the same without a second path. syncSpriteMesh now takes only the three fields of a DrawItem it reads. - The subject is resolved by previewPicture from the Shader tab, so a placed sprite shows its own overrides. - previewFit is the rule for size and place: whole device pixels per unit from 1 on, at most 8, and the corner on a whole device pixel. Pixel art stays sharp. A sprite larger than the group shrinks freely. - A shader with a Time node keeps the preview redrawing once a frame while it is on screen. A folded group or a hidden panel stops it. - New test hook __gm.inspectorSpritePreview(). Finding: the preview redrew on a model change in a microtask at first. A drawn picture whose revision moved on is retried once, in a microtask queued by the scene's synchronous redraw. That retry ran before the preview had let go, and the old pixel art sheet stayed waiting (e2e 48l-iii read waiting [1,1,1]). The preview now redraws at once on a model change, as the scene does. Tests: - unit: 238 files, 3592 tests (+5 in preview-fit.test.ts), and the group list of a placed sprite in sections.test.ts ends in PREVIEW - e2e 48: ALL PASSED, 47 checks (+2: 48l-i asset preview with the frame, 48l-iii placed preview with its shader values) - counter-checks: without whole-pixel zoom 2 unit tests fail; with the microtask redraw 48l-iii fails with waiting [1,1,1] Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
1665576fb9 |
Let a pixel perfect camera draw on the game view's pixel grid
A camera that follows smoothly stands between pixels most of the time, and pixel art then wanders between them. The new camera switch draws what the camera shows on the game view's own grid, while positions in the model keep their fractions. - `Pixel perfect` is a camera row, off by default. It works at every fixed game scale and at Auto, but not at Fit and not in the Scene tab, where one works with the fractions. - The grid has one step per device pixel of a world unit. At 100% that is whole units, at 500% fifths. The view carries it (`View.grid`), so no view draws with the grid of another. - The view looks from a point that puts the world's origin on a whole device pixel (`lookFrom`). This also covers odd camera sizes and views that cut the picture to an odd size. - Sprites, texts and tilemaps snap their top left corner when they are unturned, so an odd size lands on whole pixels too. A turned object snaps its pivot (`snapPlacement` in `pixelGrid.ts`). - Under a camera shader the picture is taken the same way, and the objects in it use the picture's grid. With Camera pixels that is one step per camera pixel. - An overlay camera has its own switch and snaps only while the main picture is at a fixed scale. - Particles are drawn where they are. - New test hook `__gm.gameDrawnRect(id)` reads where a sprite's picture is drawn, which `gameRect` (the model's) cannot tell. Found: the e2e fixture at 300% drew the crate at y 205.5 device pixels without moving the camera, because the view cut to an odd height puts its centre on half a pixel. With the switch on it is at 206. Tests: - Gate: 237 files, 3587 tests (+10 in pixel-grid.test.ts) - e2e 2,3: ALL PASSED, 75 checks (+1, 3j) - Counter-checks: without the origin snap 3 unit tests fail, with the centre snapped instead of the corner 2 fail Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
3f57cf731d |
Keep a pixel frame worn through a shader alive while it is drawn
Changing a palette colour of pixel art flooded the console with Pixi warnings: "[BindGroup] a 'textureSource' was destroyed while still bound to a shader", and the same for its sampler. #293 had stopped this for a drawing, but a frame of pixel art is worn as a cut of the sheet. The cut has a source of its own, so the sheet looked unbound, was freed at once and took the cut with it under the shader. - The check whether a replaced picture is still in use gets its key too. TextureLibrary then asks the picture and every cut taken from it. - A sheet with a bound cut waits like a bound picture did since #293, and both go together once the shader has taken the new frame. Found on the way: - The project the report came from has a sprite wearing a frame of an eleven-frame pixel art through a shader asset. - A nine-sliced sprite wearing the frame showed no warning, even without the change. The first try of the e2e check used it. Tests: - Gate on melektaus-atx: 236 files, 3577 tests (+1) - e2e 48: 45 checks (+1, sum 1741). 48l-iii puts a shader on the object wearing the second frame, changes a palette colour three times and counts the warnings. - Counter-checks bite: without the cuts in the check the unit test fails, and 48l-iii counts 6 warnings for 3 changes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
c74b6b1cab |
Show where an inspector row can be dragged and answer the pointer
After #310 rows could be dragged by their mark, but nothing said which marks take a drag, and the pointer on one changed nothing but the cursor. - A mark that drags is now a Grip: the same widget wrapped, the same size, so the value column stays where it was. Once sortBlocks takes it, it carries six faint dots at its left edge. - Marks that cannot drag stay plain: the variables of a placed object and the fields of a typed struct show no dots. - With the pointer on a grip its dots light up and the whole block that would move is washed in the accent colour, a struct with its fields. While the block is carried the wash gives way to the dimming, and the grip stays lit. - While a block is carried the other grips let the pointer pass: no lit dots, no wash. Every grip of the column shows grabbing instead of grab until the drop. - The wash first used controlHover like list rows. In the inspector it is darker than the panel and vanished next to the dark fields, so it became the accent at 15 percent, the colour of the drop line. - __gm.inspectorMark also reports the dots: none, idle or lit. Tests: - Gate on melektaus-atx: 236 files, 3576 tests (+4 in block-drag.test.ts, one of them a drag passing over another grip). - e2e 22: 40 checks (+1, sum 1740). 22f-iii reads idle dots, moves the mouse onto the mark and reads them lit. - Two counter-checks bite: no wash on hover fails 1 test, grips never made grabbable fail 3. A third: hover ignoring the drag fails 1. - Screenshots of a lit single row and a lit struct block looked at. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
2c19c89878 |
Let variables, fields, ports and list items be dragged into order (#310)
Variables of a behavior, the fields of a struct, the ports of a start or end node and the items of a list could only be added and removed. Changing their order meant removing and adding them again. Now each of them is dragged by the mark in front of its row. - sortBlocks (ui/blockDrag.ts) makes the rows of one column sortable. A block is every row it covers, so a struct moves with its fields and a list item with its nested list. The block is dimmed while it is carried, an accent line shows where it lands, and a scrolled panel follows the pointer at its edge. Less than 4 px stays a click. - The arithmetic is pure and lives in ui/sortable.ts: blockGap, gapLineY, dropSlot and edgeScroll. dropIndex of the tile strip moved there too, so the reorder maths sits in one file. - MoveVariableCommand, MoveStructFieldCommand and MovePortCommand build on the existing MoveCommand. Undo moves back. - Lists need no command. InlineList replaces the list as a whole, as it does for every other edit of its items. - The grip is the kind mark, for a list item its number. Moving only happens among siblings: a field stays in its struct. - On a Struct asset the move runs through WithStructTypes. Typed variables, Pack/Expand ports and split outputs follow the new order. - Values and edges hang on names, so only the order changes. Nodes running a behavior read their ports from the start node, the new order reaches them without a copy. - Not sortable by design: the variables of a placed object, the fields of a typed struct (moved on the Struct asset) and the picked ports of Variables/Properties nodes. - Test hook __gm.inspectorMark(label) finds the mark of a row. Tests: - Gate on melektaus-atx: 235 files, 3572 tests, CHECK EXIT 0 (+9). - tests/core/reorder.test.ts: a variable, a field two levels deep, a Struct asset reaching a typed variable and Pack, and start ports reaching the calling node with edges and result kept. - e2e 22: ALL PASSED, 39 checks (+4, sum 1739). 22f-iii drags a variable past a struct of several rows, then a field inside it. - Counter-probes: dropSlot by top edge instead of middle and a field move that ignores its path fail 3 unit tests. With sortBlocks switched off, the three order checks of 22f-iii fail. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
cf6043a066 |
Show a wheel step or Fit as a short zoom transition (#309)
The wheel and the Fit button used to jump to the new view. Now the view takes a short way there, 200 ms by default, everywhere the editor zooms. The zoom steps stay the same, only the change is shown. - viewTween: tweenView is the way between two views as a pure function. The scale moves by equal ratios, and the world point that sits at the same place on screen in both views stays there. After a wheel step that is the point under the pointer. - ViewTween drives it on the ticker. A surface hands it reading, writing and drawing its own view and wraps its wheel and Fit code in change(). A step during a transition starts from where that one was heading, so quick steps add up to exactly as many steps. - Anything else that moves the view ends the transition. A pan jumps to the target first, and saving a view saves the target. - Wired into FreeViewport (Scene, Container, animation preview), SheetCanvas (Texture, Curve, Pixel art, Vector art), GraphCanvas, panAndZoom (Tile board, plugin node board) and the timeline. - Menu > General settings... sets the time in ms, 0 jumps at once. It is kept in this browser like the panel sizes, not in the project, and it is not an undo step. - A document that opens still comes up framed at once. The animation preview frames a new subject itself instead of calling fit(). - The pixel art readout names the step the view is heading for. Measured in the scene view at 200 ms on melektaus-atx, where every frame of the transition is a full redraw: - 500 sprites: redraw 6.7 ms, 16 frames, mean 18 ms - 1000 sprites: redraw 13.0 ms, 9 frames, mean 27 ms - 2000 sprites: redraw 32.2 ms, 5 frames, mean 58 ms Tests: - The harness sets 0 ms on every load, so the existing modules read the step and not the way. - e2e 46i samples one wheel step at 1000 ms a frame later: zoom 1.033 between 1 and 2, the point under the pointer kept at (134, -40). It lands on the view a 0 ms jump reaches. - e2e 46 and 49 expect the new menu entry. - Gate (atx): 234 files, 3563 tests, CHECK EXIT 0 (+15) - e2e 46,49 (atx): ALL PASSED, 43 checks (+3, sum 1735) - Three counter-checks fail view-tween.test.ts as expected. e2e 46i fails when the scene view does not carry its centre along. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
3e05cbbed5 |
Let pixel art and vector art be nine-sliced and sorted (#308)
Nine-slice and the sort line hung on the texture frame alone. Drawing and sorting already read both off the TextureRef every picture hands out, but the generated pictures always said "no slice, no line". So the feature moves from the frame to the picture, without a second copy of any rule. - PictureCut (border, fill, sortLine) is one type. TextureFrame extends it, PixelArtAsset and VectorArtAsset carry it as a whole. The texture file format does not change, and no migration is needed. - The rules in texture.ts take a cut and a size: cutSortLine, setCutSlice, rawCutSlice and the rest replace the frame* versions. serialize.ts reads all three through one readCut. - cutTarget.ts resolves a target to cut and size: a texture frame, a pixel frame's size or the vector canvas. It lives apart from texture.ts because vectorart.ts reaches texture.ts itself. - SetSliceCommand and SetSortLineCommand take that target instead of (assetId, frameId). They refuse an asset that draws no picture. - generatedRef fills slice and sortLine from the drawing. The whole strip of a pixel art asset with several frames gets none, it is no picture of that size. With one frame asset and frame are the same. - Nine-slice scale is one shared schema row, now also on pixel art and vector art. - cutSection.ts holds the NINE-SLICE and SORTING rows once. The frame inspector and the drawing inspector both build them. - SheetCanvas draws a drawing's lines that are set, dashed as in the Texture tab. They are not dragged there, the painting tools own the canvas (as proposed in the issue). - drawOrder falls back to the picture's height when a ref has no rect, or a sort line on vector art would have been ignored. Tests: - Gate: 232 files, 3542 tests (+8: drawing-cut.test.ts, and vector art in sort-by-y.test.ts). Three counter-proofs bite: no cut on a drawing, a cut on the whole strip, no height fallback. - e2e 48: 44 checks (+1, sum 1732). 48l-ii reads red and blue where 12-pixel borders at scale 2 put them. Without borders both spots read green. - Looked at: the Pixel art tab with the lines, and the sliced sprite. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
d9c6bbfa4d |
Add a node that names the render layer of an object (#305)
The issue asked for Set render layer, Set sees and Render layer of. As with #211, a probe showed that writing already works. A Properties node of the object's type writes Render layer during a run, Inherit included, and one of type Camera writes Sees. The next picture follows both. As agreed, only Render layer of is built, and writing stays with Properties. - Render layer of is a data node with the input object and the outputs name and number. It answers the layer that holds, so an object without a choice of its own names its parent's. - NodeContext.renderLayerOf answers the number with the same rule the drawing uses (renderLayerAlong). The new field renderLayers on GraphLookup brings the names to the node. - hasRenderLayer in PropertySchema says which types are drawn. - Decided here: for an object that is not drawn (a camera, a behavior, a collider) the name is empty and the number -1, as Layer of does without a collider. Changing the render layer or Sees during a run had no test so far. Three unit tests now hold the proof from the issue. Tests: - Gate: 231 files, 3534 tests (3526 + 8), exit 0 - e2e 6: 149 checks passed, e2e 63: 13 checks passed (+1) - Unit: the coin moves into the overlay with Set and back with Inherit, a HUD container takes its sprite along, the main camera sees a second layer once its Sees is set, and Render layer of names the inherited layer, and none for a behavior - Node catalog: four cases for Render layer of - 63c3 moves a sprite onto UI in a Play run: the main camera no longer draws it, and the log reads "UI" and "1" - Three counter-probes fail 1, 1 and 2 tests: without inheriting in the node, without the check for a drawn type, and without inheriting in the drawing - "layer of" in the palette still finds Layer of first (6z-x) - The e2e sum is 1731 (1730 plus 1) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
70f3608733 |
Let Mouse position and Is hovered see through any camera (#304)
An overlay camera from #302 can stand elsewhere and zoom differently.
A HUD button only it shows lay somewhere else in the world than the
mouse of the main camera, so Is hovered never found it. Both nodes now
take a camera and see the mouse through that camera's picture.
- New input `camera` on Mouse position and Is hovered. Left open, both
read the main camera exactly as before.
- `pointerThrough` in core/properties/camera is the rule as a pure
function. It fits the camera's picture into the main one and centres
it, the same way overlayView draws it.
- The game view reports where the mouse lies on the main picture, as a
part of its width and height. The world turns that into a camera's
view, so a main camera on the move does not drag the mouse across
the HUD while the mouse stands still.
- `in view` is true only on the camera's own picture. A camera that is
not drawn (switched off, a second plain one, not a camera) never has
the mouse on it.
- On click and Mouse button stay with the main camera, as agreed in
the issue.
Also fixes a unit test that
|
||
|
adminandClaude Opus 5.5
|
2aa635f453 |
Keep an output's name readable beside a long inline value (#306)
An input and an output in the same port row were cut to fit each on its own. The input text could reach x = 123 and the output label could start at x = 83, so a long inline value ran under the output's name. At Animation playing a long animation name covered "progress". - The output label is placed first and keeps its room. The input text is cut with an ellipsis 6 px before it, as decided in the issue. - The rule is a pure function in src/editor/graph/portRow.ts. A row without an output keeps the input's old limit of 108 px. - The value a run shows beside an output uses the same rule. Its left edge can no longer lie right of its right edge. - New test hook __gm.graph.rowSpans(nodeId) gives each row's input and output text as cut and placed, in node units. - NODE_WIDTH stays as it is, as decided in the issue. Tests: - Gate: 230 files, 3520 tests (6 new in port-row.test.ts). - e2e 6: all passed, 149 checks (148 + 1). - Counter-check: with the rule switched off, the input text reaches x = 121.4 under "progress" at x = 114.1, and the new check fails. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
808e1eefc2 |
Clean up duplicates and dead code, and fix the + menu check (#85)
Refactoring round 9 covers everything since round 8 ( |
||
|
adminandClaude Opus 5.5
|
a526e1b40f |
Use a shader graph as a node in other shaders (#303)
A calculation needed in several shaders (a greyscale, a noise, an outline) had to be built again in every one of them. A shader function keeps it as an asset of its own, called like a behavior is since #16. - New asset type `shaderfunction` ("Shader function"), edited in the Shader tab. The panel beside the graph shows no picture, the readout counts its inputs and outputs. - Two new nodes: `Function input` and `Function output`. Their declared ports, numbers or colours only, are the ports of every call. One of each per function, like Start and End. - The call is node `ShaderCall`, named after the function. The palette lists every function, it can be dragged in from the Assets panel, and a double-click opens it. - The compiler sets the function in where it is called (`inlineFunctions`). It renames the copied nodes `<call>/<node>` and turns an unconnected port into a constant. Passes, Time, the Edge cut and the corner walk then see a plain graph. - A problem inside a function is reported at the call, in the function's words. The function's own tab checks it on its own (`checkShaderFunction`). - `ShaderLibrary` keys a shader on its graph and on every function it calls, however deep, so a change reaches every caller. - The port commands mirror a new, renamed or removed port into every call, as for a behavior's Start and End. `NodeDef.portKinds` holds the two ends to Number and Color. - As agreed in the issue: no `Property`, `Color property`, Output or Vertex output in a function, and `Time` is the caller's clock. - Own decision: a `Pass` may stand in a function. It becomes a pass of the calling shader, so a wide blur can be a function. - A function calling itself, also through another, says so at the node and is not set in. The palette does not offer such a call. - The Shader tab showed only the compiler's problems. A call nothing reaches was never compiled and looked healthy, so a call now also shows its own problem there, as in a behavior. Tests: - Gate: 228 files, 3510 tests, CHECK EXIT 0 (+13: 11 in tests/core/shader-functions.test.ts, 2 for the new icon) - e2e 28: 100 checks, all passed (+3 in 28l-viii). Called and in place both #606060/#a2a2a2, halved in the function #303030/#515151. - Counter-check: a fingerprint without the called functions fails 1 unit test, and in the browser the change check (#606060 stays). - The first e2e runs found two faults, both fixed: a double-click only selected the function, and an unreached call of itself showed no problem. - e2e total 1728 is computed (1725 + 3). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
2e55071fa5 |
Lay a shader over a camera's whole picture (#299)
A shader so far worked only on its own wearer. Effects over the whole game picture, such as a vignette, grey, scanlines or a red flash on a hit, had no place to live. A camera now wears one, drawn the way a field shader of a particle source is (#170). - The camera gets a Shader row and, while one is picked, a Shader resolution row. The shader's properties follow as camera rows. They load, animate and appear in a Properties node under Type Camera. - Every view keeps masks and objects in one container. CameraShade renders it into the camera's picture, cleared to the background, and draws the shader over the frame. Game tab and play.html take the container off the stage while the picture stands in for it. - Screen pixels gives the picture as many pixels as it covers on screen. Camera pixels uses View width x View height and scales up without smoothing. - A pixel counts as a camera pixel in both (the user's choice). The uniform uPixelScale makes Texture size read View width x View height, so a Blur keeps its reach at any game scale. - Each camera wears its own shader, overlay cameras included. A red flash on the game camera leaves the HUD of an overlay camera alone. - The scene view's eye has a Shader switch on the Camera row, off by default. With it on, the main camera's shader lies over its frame, below symbols and gizmos, and the world around it stays drawn. - New tool tools/perf-camera-shader.mjs measures the cost. Measured on stein2 (HD 630, Game tab at Fit 1310x679): a camera shader that hands the picture through stays inside the scene's own spread of 3 to 5 fps between runs. A Blur of 4 cost 4 to 8 fps. Tests: - Gate on melektaus-atx: 227 files, 3497 tests (3487 + 10). - e2e 63 alone: 11 checks (6 + 5), e2e sum 1725 counted. - 63e-63i: pass-through is byte-equal and upright at 100%, flat red reaches the corners while the overlay keeps its green, Texture size reads 640 x 360 in both resolutions at Fit, no shader draws as before, and the scene view shows it only with the switch on. - Counter-checks: a camera left out of the shader wearers fails 1 unit test. Counting screen pixels fails 63g with #ab6000. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
b2326085ed |
Show each camera only its layers and lay overlay cameras on top (#302)
A camera could only show the whole scene, so a HUD drawn in the world would run through a camera shader (#299) with everything else. Render layers say which cameras see an object, and overlay cameras draw their own picture over the main one. - The project has a list of up to 16 render layers, edited in Project settings next to the collision layers. It is a list of its own, read by the same rules (`layerNamesOf`). - Sprites, texts, tilemaps, particle sources and containers carry a row `Render layer` under their switches. Its default `Inherit` takes the parent's layer, at the top the first one. - A camera has `Sees` (all layers by default) and `Overlay`. The main camera is the first one switched on without Overlay, so two plain cameras still draw through the first one only. - Every view through a camera keeps only what that camera sees and lays one see-through view per overlay camera over itself. That view is a `WorldViewport` of its own, handed its view and frame from below. - An overlay is fitted into the main camera's whole picture without distortion. One of the same size and zoom covers it exactly. - `z` sorts within one camera only. The pictures lie on top of each other as a whole, which is what a shader per camera (#299) needs. - The Game tab's preview now collects the scene once per redraw. Before it collected it twice, with the cameras it would have been five times. - `SetCollisionLayersCommand` became `SetLayerNamesCommand`, which takes the list it sets. Follow-ups: #304 (the mouse through a chosen camera) and #305 (change the render layer at run time). Tests: - Gate: 226 files, 3487 tests (3480 + 7) - e2e 63 (new): 4 checks. A green overlay sprite at z -10 covers a red main sprite at z 1000, in the preview and in the running game. A `UI` sprite in front of the main camera and a `Default` one in front of the overlay stay away. Without Overlay the second camera draws nothing. The settings dialog adds a render layer at the end. - Counter-checks: inheriting from parents switched off fails 1 unit test. Overlays never drawn fails the two overlay checks in e2e 63. - e2e total 1720 (1716 + 4), computed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
f03ca0d382 |
Draw a shader in passes with a Pass node (#301)
A Blur stops at radius 8 in one pass, and bloom needs a softened copy of the picture beside the original. A Pass node cuts the graph: what feeds it is drawn into a picture of its own first, and the nodes that read a picture read that one when it feeds them. - Pass has color and alpha like the Output, a Resolution setting (Full, Half, Quarter) and a picture output. Texture color, Texture alpha, Sample texture, Blur and Texture size get a picture input. Unconnected it is the wearer's own picture, so the original stays readable beside any pass. - picture is a port type of its own. It plugs only into a picture input, not even into an any port. - The walk shared by the compilers can stop at a node (cut). The shader compiler draws each Pass it meets as a fragment of its own, earlier passes first. A ring through a picture says Feeds itself, and more than 8 passes say so at the node. - A pass draws untinted. The tint comes on when the last part reads a picture, so a particle source draws its passes once for all of its particles (user decision). - A pass's picture is the wearer's picture and its Padding, at the pixels of the picture times Resolution, at most 2048 per side. It is read with the picture's own filter and continues past its frame as the Output's Edge says. - PassPictures keeps one render texture per pass and wearer. It draws again only when picture, size, Padding or a uniform changed. A field's picture is a render texture that changes under the same id, so a field draws its passes every time. - Sprite, sliced sprite, text, particle source, field shader and the Shader tab's preview all draw passes. syncShader takes a fixed picture for a 9-slice, which holds its corners but keeps the margin. - texelAt and blur take the sampler they read. Existing shaders compile to the same program apart from that argument. - A field shader's Blur and Texture size count a pixel as a unit of the frame (uPixelScale), not of the screen. Before, a field's blur grew when zooming out and shrank when zooming in. This came with #170 and #300 and is fixed here because passes inherit it. - tools/perf-shaders.mjs and tools/perf-field-shaders.mjs measure two passes with GM_PERF_PASSES=1 and GM_PERF_PASS_RESOLUTION. Below Full an open Pass shrinks the picture first, as the wiki's recipe does. Found on the way: - Destroying the pictures before the shader made Pixi warn that a texture was still bound. The shader goes first now, and a picture of a new size is resized rather than replaced. - Two passes in a row could draw plain white the first time. Pixi builds the code that binds a program's resources once and finds them by position. A pass of one graph is often the same program as a pass of another with fewer slots, so it bound a placeholder. Every shader now declares all 8 pass slots and its uniform group. Tests: - Gate on melektaus-atx: 225 files, 3480 tests (3467 + 13), exit 0. - e2e 28 and 31 on melektaus-atx: 97 checks (94 + 3) and 44 checks (43 + 1), all passed, no Pixi warning in the console. - 28l-vii: a Blur across, a Pass and a Blur down draw the 5x5 spot of a Blur of Both (difference 0.000). On a grey sprite a pass's picture added to the original is 231/60/60, each tinted once. Two Passes in a row after a graph with one show red and green, not white. - 31v-iii: a Blur of 8 fades a field's frame edge over 8.0 units at 50 % and at 200 % zoom. - Counter-checks: a tinted pass fails 1 unit test. Pictures never bound fail both e2e checks (spot 0.20 instead of 0.04, 243/158/158). Pass slots by the number of passes draw #ffffff, and a field without its pixel scale ramps over 16 and 4 units: exactly the two new checks fail. - e2e sum 1716 is counted, not yet measured (1712 + 4). - Measured on stein2 (HD 630, one run per row): 50 field shaders at radius 8 go from 11.4 to 41.3 fps in two passes, 44.4 at radius 16 (three passes at Half). 800 sprites at radius 8 go from 13.9 to 24.5 fps. At radius 4 one pass stays faster from 400 sprites on. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
df61b08bd1 |
Soften a shader's picture in one pass with a Blur node (#300)
A blur already worked with Sample texture, but only by wiring every sample by hand: 25 nodes for a 5x5. Averaging their colours also gave a dark seam wherever a neighbour was transparent. The Blur node writes the loop itself and averages the right way. - Blur takes radius (pixels of the picture, 0 to 8, fractions allowed) and u/v like Sample texture, and gives color and alpha. - Kind is Box (default) or Gaussian, Direction is Both, Horizontal or Vertical. Both are arguments of one prelude helper, blur(). - The samples are averaged premultiplied and only then made straight, so a transparent neighbour lowers the opacity and adds no black. - The outermost ring counts by the part of the radius that reaches it, so a radius from a Property changes smoothly and compiles nothing again. - A node may now declare prepare, a vec4 the compiler works out once into a local. Color and alpha read that one sum instead of blurring twice. - A reached Blur counts as a shifted sample, so the picture comes with the Output's Edge. It cannot feed the Vertex output. - Both perf tools take GM_PERF_BLUR=<radius>. Wider blurs and bloom need several passes. That is #301. Measured on stein2 (HD 630, one run per value, fps), Box blur: - 48x48 sprites at 100/400/800: radius 2 59.8/58.4/55.6, radius 4 53.1/51.4/36.3, radius 8 46.6/24.4/14.4. - 256x256 field shaders at 1/10/50: radius 2 60.2/56.0/49.0, radius 4 60.1/49.0/27.1, radius 8 52.4/34.9/11.4. Tests: - Gate: 224 files, 3467 tests, CHECK EXIT 0 (+7). - e2e 28: ALL PASSED, 94 checks (+5 in 28l-vi). Pixels match the arithmetic to 1: the last red texel at radius 0.5/1/4 reads 187/173/154 (expected 188/173/154). - Counter-checks: Blur not counted as shifted fails 1 unit test. Blur output left premultiplied fails the seam and spot checks. - e2e total 1712 is computed (1707 + 5). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
d6d38a6422 |
Lay a shader over a particle source's whole field (#170)
A shader on a source runs per particle and stops at each quad, so a glow or blur across the whole cloud was not possible. A source now has a second shader row that is laid over everything it draws at once. - New rows Field shader, Field width and Field height (256) on a particle source. The size rows show only while a field shader is picked (new `shownWhenSet` on a row definition). - The particles are drawn into a RenderTexture of the frame, and a sprite mesh wearing the field shader draws that picture. The frame is centred on the source and turns and grows with it. - UV run 0 to 1 across the frame. Texture color, Texture alpha and Sample texture read the picture of the particles. - The picture has as many device pixels as the frame covers on screen, at most 2048 per side. A frame off screen draws no pass. - The field shader's Padding and Vertex output do nothing: the frame is the edge, and its corners stay put as on a nine-sliced sprite. - A name both shaders declare is one row, and both uniforms read it. - The scene view draws the frame dashed round a selected source. The Particles tab preview wears the field shader too. - `tools/perf-field-shaders.mjs` measures 1, 10 and 50 sources. On the HD 630 they ran at 60.0, 60.0 and 57.6 fps against 60.1, 60.0 and 59.2 without, one run per row. Tests: - Gate: 223 files, 3460 tests (3452 + 8), CHECK EXIT 0. - e2e 31: 43 checks, 3 of them new. Sum 1707 is computed (1704 + 3). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
bf27e42e8a |
Let a shader declare a color the sprite sets (#298)
A shader could only ask its sprites for numbers. A tint, an outline colour or a glow colour had to be built from three Property nodes. - New node Color property next to Property: Name and Default (a colour, white unless picked), output `color`. It becomes a colour row below the Shader dropdown and a vec3 uniform in the program. - Number and colour properties share one set of names. A clash is the existing "Another property is named" problem. - Like Color it cannot feed the Vertex output. - Placed from the palette it is called tint, then tint2. - The preview in the Shader tab shows a colour field for it. Like the sliders it starts at the sprite's value and never writes into the project. - `declaredRows` takes the row builder, so the particle graph keeps its number rows unchanged. Tests: - Gate: 222 files, 3452 tests (3445 + 7) - e2e 28: 89 checks (28l-v: default colour drawn, then the colour the sprite sets) - e2e total 1704, counted (1703 + 1) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
4f4c149211 |
Duplicate an asset from its right-click menu (#297)
A new asset that starts out like an existing one (a death animation
next to "Break", a second slime) had to be built again by hand. Now
the asset panel's right-click menu has Duplicate.
- The copy lands right below the original and is named the way a new
asset would be ("Crate 2"). It is selected and its tab opens where
it has one. A copied scene does not become the active one.
- One undo step takes it out again. A redo brings back the same id.
- The copy is a deep copy: values, graph, tracks, layers and objects
are its own, and editing it leaves the original alone.
- Frames of a texture or of pixel art get fresh ids, because a sprite
names a frame across the whole document.
- Objects of a scene or container get fresh ids the way pasted ones
do. Parents and object references among them follow the copy.
- Whatever pointed at the original keeps pointing at the original: a
sprite wearing it, an animation's root, a placed object.
Tests:
- 3 new unit tests (DuplicateAssetCommand). One counter-check run:
without renaming the references the scene test fails.
- e2e 04: 2 new checks with the real menu (copy named "Crate 2" right
behind the Crate and selected, one undo removes it).
- Gate: 221 files, 3445 tests, CHECK EXIT 0.
- e2e total 1703 is computed (1701 + 2). Full suite left to CI.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
adminandClaude Opus 5.5
|
f1e3bd8378 |
Let a joint glide on a rail and drive it with a motor (#154)
Step 2 of the joints: a door that opens by itself, a lift and a piston need a joint that slides along one line and a motor that turns or pushes. A graph also needs to read how far such a joint stands. - Slider: a fifth joint type on Rapier's prismatic joint. Its rail is the joint's own x axis, so turning the joint turns the rail. Joints now carry Scale, Rotation and Mirror like a sprite. - Limit distance with Min and Max, in world units from where the object was placed. - Motor on hinge and slider: Off, Speed or Target, with Motor speed, Motor target and Max force (0 = no limit). Target goes there at no more than the speed and slows down for the last stretch. - Changing speed, target or force keeps the engine joint. Switching the motor on or off builds it again. - New node Joint position: value and speed of a hinge (degrees, counting past a full round) or a slider (world units). - Scene view draws a slider's rail dashed, with a tick at each limit. - shownFor in the schema takes a list of choices that must all hold. Findings, measured with a probe against Rapier 0.20: - A prismatic joint holds both bodies at one angle. Held to the scene or to an object without a rigidbody, the joint's own anchor body is turned by the placed difference, so nothing jumps. Between two rigidbodies the held one is turned to match, and the console says so once. - Rapier reads no joint angle back. The position is worked out from the two bodies. - A hinge motor is weak on a small body at the end of a long arm (box 20 on an arm of 100 does not lift itself). More solver iterations would help but cost every scene. A door of 100 x 20 reaches -90 and holds it. This is written in the wiki. Tests: - Unit: 16 new in joints.test.ts, 1 in joint-links, 3 catalog cases. - e2e 41: 27 checks (25 + 2), block 41f with a real mouse in the inspector and the motor in play (59.1 of 60 units in 1 s). - 9 of 9 unit counter-probes fail as they should. - Gate: 220 files, 3436 tests (3416 + 20), CHECK EXIT 0. - e2e total 1701 is counted (1699 + 2), the full suite did not run. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
4716726b6b |
Let a shader move the corners of the picture it wears (#79)
A shader could only say what colour a pixel has. Where the picture is drawn was fixed to the four corners of its rectangle, so nothing could wave, bend or wobble. - The new node Vertex output takes x and y, in pixels of the picture, and moves every corner of the wearer's grid by them. The UVs stay with the corners, so the picture bends along. - Subdivisions on that node says how fine the grid is: cells per axis, 1 to 64, 8 by default. 1 is the plain quad. A shader without the node keeps the quad every sprite shares. - A graph holds one Vertex output at most. UV, Time, Property, Number and the math nodes may feed it, and one node may feed both outputs. - Texture color, Texture alpha, Sample texture, Texture size and Color cannot feed it. The node says "Not available per corner" and the shader does not compile. - Sprites, texts and the preview in the shader tab wear the grid. A nine-slice sprite and particles draw as if the node were not there. - Picking, the selection frame, the size and colliders stay with the rectangle. A wearer with a Vertex output is never culled, as with Padding, and with Padding the grid covers the room around the picture too. - tools/perf-shaders.mjs takes GM_PERF_SUBDIVISIONS to measure a grid against the plain quad. Measured on stein2 (Intel HD 630), frames per second, one run per row: at 800 sprites 58 with a shader on the quad, 46 with 8 cells, 41 with 32. Up to 400 sprites a grid hardly shows. Every sprite holds its own grid. Tests: - Gate on melektaus-atx: 220 files, 3416 tests, CHECK EXIT 0. - 15 new unit tests for the compiler and the grid. - Four new e2e checks in modules 28 and 30: a fixed offset moves the picture, 8 cells bend only the lower half where 1 cell shears it all, Texture alpha is refused at the node, a text moves. - e2e 28,30 on melektaus-atx: 110 checks, all pass. - The e2e total of 1699 is counted (1695 + 4), the full suite did not run. - No counter-probe, no screenshot. Not looked at: the preview, Time in the browser, flipped sprites, Padding together with the grid. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
963f8a152a |
Open a scene framed, and thin out two inspector sections (#295)
Three small things asked for in one issue. A scene came up wherever the view happened to stand, and the inspector carried a button and a list that are better reached somewhere else. - A scene comes up framed as if Fit had been clicked: after a project is loaded and when another scene is opened. From then on the view belongs to the user. - An empty scene has nothing to frame and stands at the origin, one pixel per unit. A scene the running game loads is not framed anew. - The Container tab is left out. It keeps the view of each container as before (#171). - The Unpack button and the two lines under it are gone from the inspector of a placed container. Unpack container stays in the right-click menu of the hierarchy and of the scene view. - The corners of a polygon collider are no row in the inspector any more. They are edited on the outline in the scene view. The new schema option noRow keeps a property out of the inspector while it stays a property. Tests: - Gate on melektaus-atx: 219 files, 3401 tests, CHECK EXIT 0. - One new unit test in tests/core/polygon.test.ts: the corners are a property of the collider and no row. - Three new e2e checks: a loaded project is framed like Fit (module 03), an opened scene is (04), and after a real reload a click on Fit moves nothing (09). - Module 09 unpacks through the right-click menu now and checks that the inspector offers no button. - Full suite on melektaus-atx: 1695 checks, all pass. Smokes 24, 19 and 21, all pass. - No counter-probe evaluated, and nothing looked at after the build. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
adminandClaude Opus 5.5
|
ca5b2db449 |
Keep where each divider was dragged to across a reload
A panel dragged wider was back at its first size after every reload. The sizes are now kept in the browser, next to the folded groups. - editor/panelSizes.ts holds one entry in localStorage, game-maker.panel-sizes: a name per divider and its width, height or share. - A panel reads its size once when it is built, held to the bounds its divider keeps today. Every drag writes the new size. - Kept are the columns left and right of the centre, the console, the split of the left column, the tilemap panel, the preview of the Shader, Particles and Pixel art tabs, each under its own name, and the split of the user management. - The sizes belong to the browser, not to the project. They hold for every project opened there. - A browser that refuses the storage, or holds rubbish under the entry, starts with the sizes the editor comes with. Tests: - Gate on melektaus-atx: 219 files, 3400 tests, CHECK EXIT 0. - Seven new unit tests in tests/editor/panel-sizes.test.ts. All five counter-probes bite. - Smokes on melektaus-atx: 24, 19 and 21, all pass. - One new e2e check in module 48: the left column dragged 40 px wider is as wide after a reload, and dragged back after. - Full suite on melektaus-atx: 1692 checks, all pass. It ran before the fix to the animation players of the commit before last. - Browser counter-probe: without the write to the browser exactly that check fails. - Only the left column is measured across a reload. The other dividers go the same way and have no check of their own. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |