* feat(studio,cli): per-frame board comments, self-refreshing storyboard, status-aware preview landing Per-frame comment boxes on the storyboard board batch into .hyperframes/frame-comments.json (a resubmit wins per frame; unconsumed comments on other frames are kept). Submitted-but-unconsumed comments stay visible — a toolbar banner plus a per-tile echo — until the agent consumes the file; the banner also says what to do next (reply anything in the agent chat). The board keeps itself current: GET /projects/:id/signature exposes the watcher-cached project signature, the storyboard payload carries the signature it was derived from, and the view polls at 2s (hidden tabs skipped, re-checked on visibility), refetching in place with no loading flash. Posters bake the signature into their URL so tiles fill in as sketches land and a poster that failed mid-write retries on the next version; the empty state upgrades itself when STORYBOARD.md appears, and its handoff prompt now points the agent at the review loop and uses the parser's real status vocabulary (outline, not planned). preview lands the browser on the storyboard view while the board is the review surface — any frame built, or pure planning (srcs declared, none on disk yet) — and on the timeline once the video is assembled. * feat(skills): the review loop — plan, sketch, build as one shared process hyperframes-core/references/review-loop.md is the single source for the three-pass collaborative review: the plan proposed on a live board (§ 1), wireframe sketches marked built with one layout question (§ 2 — real words on plain blocks, run no CLI; a confirmed board is itself a valid deliverable when the user asked for a storyboard, not a video), the build dressing confirmed layouts (§ 3, worker or inline), and the final look (§ 4). Autonomous runs skip every gate and keep one question before render. The three narrative workflows' Steps 3/4/6 collapse to references plus their sketch stand-ins (captured-asset blocks for product-launch-video, plain code panels for pr-to-video); the confirmed-sketch handoff stays in each frame-worker prompt. general-video plans on a board for multi-scene narrative pieces in collaborative mode — its sketch pass is layout-before-animation with the user watching. The router treats "I want a storyboard" as a process request rather than a route, and closes exploratory intake by recommending a route plus how the run will review. The supporting contracts land next door: the comments channel (silent submit, one reply picks it up, check the file before the words) in brief-contract § 1; the sidecar schema and the built status rung in storyboard-format; the mode question asked first and alone in the three workflows' Step 0. * feat(media-use): user memory — remembered preferences and frozen recipes Two tiers of memory on media-use's existing two-tier storage split. Preferences (lightweight): confirmed brief answers — destination, aspect, language, mode, voice, style preset — recorded to the project's .media/preferences.json (committed, the team inherits it) and promoted to the personal ~/.media/preferences.json once the same value is confirmed in two different projects (a sightings ledger accumulates the cross-project evidence user-side, since project files can't see each other). prefs.mjs get/record; merge reads project-over-user; a changed value restarts its provenance. Recipes (heavyweight): one approved run frozen as a named, versioned bundle — frame.md, the storyboard skeleton (structure kept: durations, transitions, srcs, Video direction; statuses reset to outline; content blanked to per-frame fill-ins naming the beat's role), and the confirmed brief values. Named folders, not content hashes: re-freezing bumps version and archives <name>@v<N>; a freeze is already confirmed, so it promotes to the user tier immediately. recipe.mjs freeze/list/use, plus resolve --type recipe --entity <name> delegating like grade/lut. 16 new node --test cases; the media-use lib suite is 168/168. * feat(skills): wire user memory into the brief and the review loop brief-contract § 2 gains Remembered defaults: read the merged preferences before Round 2 and let a remembered value become the recommended option with a receipt naming its source project. Memory changes the default, never the question — every ask-marked field still gets asked, and what the request says this time beats what was picked last time. Record only what the user actually confirmed (a defaulted voice nobody chose is not an answer; a "go" that accepts the recommended defaults is). The first record announces itself once; after that the receipts carry the reminder. In autonomous mode a remembered value becomes the decided value, receipt included. The three narrative workflows read the remembered defaults before Round 2, record the confirmed answers at the Step 0 gate, record the chosen preset at the Step 2 gate (pr-to-video excepted — its preset is fixed), and fall back to the remembered voice when the request names none. general-video's discovery reads the same defaults. Recipes wire in at both ends: Step 0 checks for a matching recipe before the mode question — one question, plural-aware, and adopting one fills the brief, skips the design step, and drafts the storyboard from the frozen skeleton while every review gate still runs. The review loop's final look (§ 4) offers the freeze once after approval, and the confirmation teaches the recall phrase — the name is something the system reminds the user of, never something they must remember. The router recognizes a named recipe or "like last time" as a route. * docs(skills): the sketch pass names check, not the deprecated validate * feat(skills): intent-layer references — process, route briefs, capability menu, BRIEF.md format * feat(media-use): brief skeleton as the recipe's fourth artifact; flow/storyboard preference keys * feat(skills): the intent layer conducts every brief — workflows execute BRIEF.md * feat(skills): retire the mode preference key; sync catalog surfaces for intent layer * refactor(skills): dedupe router vs intent-layer guidance — one owner per rule * feat(skills): the design ask — own spec, pick by eye from showcases, or defer * docs(skills): the design ask says the honest line on capture routes * feat(skills): product-launch-video absorbs website-to-video as the tour angle * refactor(skills): keep product-launch-video pristine — a tour is brief intent, not a pipeline branch * feat(skills): production loop + genre lenses; general-video goes freeform (route yours, laws hold) * refactor(skills): /hyperframes is the front door - route tables and scope lists leave the workflows * docs(skills): review-loop pass across skill catalog * fix(cli): pass project dir to openStudioBrowser in background-server path * feat(skills): add pitch-round reference - verbalized sampling concept gate * feat(skills): wire pitch round into intent layer - completeness triage + route eligibility * feat(skills): editorial capability recommendations, handoff disciplines, menu-probe split * feat(skills): pitches carry their machinery; source-only-formed requests pitch the telling * feat(skills): companion goes director - ceiling treatment plus blueprint/rule citation discipline * fix(scripts): sandbox npx-leak guard - private npm global prefix keeps npx on the branch CLI * chore(skills): resync manifest hash after formatter pass reflowed general-video tables * fix(skills): recipe freeze reads workflow from BRIEF.md; style_preset records require workflow scope Two holes found by a live companion-run freeze: the agent-supplied --workflow contradicted the run's actual workflow (recipe.json said faceless-explainer, brief-skeleton said general-video), and the style_preset lookup missed because the preference had been recorded under the bare key. - freezeRecipe resolves the workflow from BRIEF.md frontmatter; the flag is a fallback for briefless projects and a contradicting flag is ignored (noted). - recordPreference refuses a bare style_preset — the scoped key is the only writable shape; freeze tolerates legacy bare records via read fallback. - review-loop § 4 / media-use SKILL / brief-format wording follow the machinery.
18 KiB
name, description, metadata
| name | description | metadata | ||
|---|---|---|---|---|
| motion-blur-streak | Fake directional velocity blur on a fast entrance or camera push-through — blur peaks at max speed and resolves to 0 at the settle, so the element streaks in then snaps sharp. Two paths — SVG feGaussianBlur on the motion axis, or an echo/ghost trail that collapses into the lead. |
|
Motion-Blur Streak
Real per-frame motion blur isn't available to a seeked renderer (it integrates over shutter time, which a paused timeline has no concept of), so this rule fakes it for a fast element fly-in or a hard camera push-through. The blur peaks at maximum velocity and resolves to 0 at the settle — the element reads as streaking in then snapping sharp on arrival. The whole point is the coupling: the blur envelope rides the same ease and window as the position tween, so peak-blur lands exactly on peak-speed, and the element is razor-sharp the instant it stops.
Two implementation paths, both finite, deterministic, and seek-safe:
- (A) Directional SVG blur — an inline
<filter>with<feGaussianBlur stdDeviation="X 0">(X on the axis of motion, 0 across it). GSAP tweensXfrom high → 0 through a proxy that callssetAttributeeach frame. Cleanest, one element, true directional smear. - (B) Echo / ghost trail — 2–4 duplicate copies at decreasing opacity, offset backward along the motion vector, collapsing into the lead element as it settles. No filter cost; reads as a "speed-line" stutter trail. Better when you want the streak colored or stylized rather than a literal optical blur.
This is for entrances and mid-shot moves only — a fast arrival, a beat that zooms past camera, a card slamming into a grid slot, a logo punching through into a lockup. Never an exit on a non-final frame (a blurred element fleeing off-frame mid-composition reads as a glitch, and a hard exit between scenes is the transition's job, not a per-element blur).
How It Works
A fast move has a velocity profile: it accelerates off the start, peaks, then decelerates into the settle. An out ease (expo.out, power4.out) front-loads that — velocity is highest right at the start and bleeds to zero at the end. The fake works by mapping a blur (or echo) envelope onto that same curve:
- Position tween — the element travels from an off-frame / pushed-back start to its resting transform (
x/yfor a fly-in,scalefor a push-through) on a fastoutease overMOVE_DUR. - Blur envelope — in lockstep over the same window and ease, the smear goes from
PEAK_BLUR→0. Because the ease front-loads velocity and the envelope shares it, max blur coincides with max speed, and blur hits exactly0as the element lands. - Settle is sharp — by
MOVE_START + MOVE_DURthe element is at its resting transform with blur0(path A) or all echoes collapsed onto the lead (path B). It then holds, fully crisp, for the climax dwell.
Path A tweens the filter's stdDeviation attribute (a non-DOM-style numeric attribute) via a proxy object — GSAP can't tween an SVG attribute directly, so you tween a plain { v: PEAK_BLUR } and write it back with setAttribute in onUpdate. Path B places the ghosts at deterministic backward offsets (i * ECHO_STEP_PX) and fades/collapses them on the same envelope.
HTML
Path A — directional SVG blur (recommended default)
<div
class="scene"
id="streak-scene"
data-composition-id="streak-scene"
data-start="0"
data-duration="DURATION"
data-track-index="0"
>
<!-- Inline filter: blur on the X axis only (stdDeviation="X 0"); 0 on Y keeps it a clean horizontal smear. -->
<svg width="0" height="0" aria-hidden="true" style="position: absolute">
<defs>
<filter id="streak" x="-50%" y="-50%" width="200%" height="200%">
<feGaussianBlur id="streak-blur" in="SourceGraphic" stdDeviation="0 0" />
</filter>
</defs>
</svg>
<div class="streak-stage">
<div class="streak-el" id="streak-el">{phrase}</div>
</div>
</div>
Path B — echo / ghost trail
<div
class="scene"
id="streak-scene"
data-composition-id="streak-scene"
data-start="0"
data-duration="DURATION"
data-track-index="0"
>
<div class="streak-stage">
<!-- N-1 ghosts BEHIND the lead, then the lead on top. Ghosts are aria-hidden duplicates. -->
<div class="streak-ghost" data-i="3" aria-hidden="true">{phrase}</div>
<div class="streak-ghost" data-i="2" aria-hidden="true">{phrase}</div>
<div class="streak-ghost" data-i="1" aria-hidden="true">{phrase}</div>
<div class="streak-el" id="streak-el">{phrase}</div>
</div>
</div>
CSS
.scene {
position: relative;
width: 100%;
height: 100%;
display: grid;
place-items: center;
background: {sceneBg};
font-family: {font};
overflow: hidden; /* the smear/echo extends past the resting position before settling */
}
.streak-stage {
position: relative;
display: grid;
place-items: center;
}
.streak-el {
position: relative;
z-index: 2;
font-size: EL_FONT_SIZE;
font-weight: 900;
letter-spacing: EL_TRACKING;
color: {textColor};
/* Path A only — reference the directional filter. (Omit for Path B.) */
filter: url(#streak);
will-change: transform, filter;
}
/* Path B ghosts — identical glyphs behind the lead, decreasing opacity */
.streak-ghost {
position: absolute;
inset: 0;
display: grid;
place-items: center;
z-index: 1;
font-size: EL_FONT_SIZE;
font-weight: 900;
letter-spacing: EL_TRACKING;
color: {textColor};
opacity: 0;
will-change: transform, opacity;
pointer-events: none;
}
GSAP Timeline
Path A — directional SVG blur
<script src="https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js"></script>
<script>
window.__timelines = window.__timelines || {};
const tl = gsap.timeline({ paused: true });
// GSAP can't tween an SVG attribute directly — tween a proxy and write it back each frame.
const blurNode = document.getElementById("streak-blur");
const blurProxy = { v: PEAK_BLUR };
const writeBlur = () => blurNode.setAttribute("stdDeviation", `${blurProxy.v} 0`); // X-axis only
writeBlur(); // seed frame 0 so a seek to t=0 shows the streaked start, not a sharp pre-frame
// Position: travel from off-frame to rest on a fast `out` ease (velocity front-loaded).
tl.fromTo(
"#streak-el",
{ x: ENTER_FROM_X, opacity: 0 },
{ x: 0, opacity: 1, duration: MOVE_DUR, ease: MOVE_EASE },
MOVE_START,
);
// Blur envelope: SAME window + SAME ease so peak-blur == peak-speed, resolving to 0 at the settle.
tl.to(blurProxy, { v: 0, duration: MOVE_DUR, ease: MOVE_EASE, onUpdate: writeBlur }, MOVE_START);
window.__timelines["streak-scene"] = tl;
</script>
Path B — echo / ghost trail
<script src="https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js"></script>
<script>
window.__timelines = window.__timelines || {};
const tl = gsap.timeline({ paused: true });
// Lead element: same fast `out` move as Path A.
tl.fromTo(
"#streak-el",
{ x: ENTER_FROM_X, opacity: 0 },
{ x: 0, opacity: 1, duration: MOVE_DUR, ease: MOVE_EASE },
MOVE_START,
);
// Ghosts: each starts FURTHER back along the motion vector (deterministic, by index),
// dimmer, and collapses onto the lead — all on the SAME ease/window so they vanish at the settle.
const ghosts = gsap.utils.toArray(".streak-ghost");
ghosts.forEach((g) => {
const i = Number(g.dataset.i); // 1..N-1, set in HTML (no Math.random)
tl.fromTo(
g,
{ x: ENTER_FROM_X - i * ECHO_STEP_PX, opacity: GHOST_BASE_OPACITY / i },
{ x: 0, opacity: 0, duration: MOVE_DUR, ease: MOVE_EASE },
MOVE_START,
);
});
window.__timelines["streak-scene"] = tl;
</script>
Variations
Vertical streak (rise / drop-in)
Swap the motion axis: use y instead of x for the position tween, stdDeviation="0 Y" for Path A (blur on Y, 0 on X), and ENTER_FROM_Y / vertical echo offsets for Path B. A phrase that streaks up into place pairs with kinetic-type-beats' rise-rotate beat.
Camera push-through (scale streak into a lockup)
Instead of translating, the element rushes the camera: scale: SCALE_FROM → 1 on the fast out ease, with a radial / zoom blur feel approximated by a symmetric stdDeviation="B B" envelope (blur on both axes since the smear is depth-wise, not directional). This is the logo-assemble-lockup push-through — the wordmark punches forward out of soft focus and snaps crisp at the lock.
tl.fromTo(
"#streak-el",
{ scale: SCALE_FROM, opacity: 0 },
{ scale: 1, opacity: 1, duration: MOVE_DUR, ease: MOVE_EASE },
MOVE_START,
);
tl.to(
blurProxy,
{
v: 0,
duration: MOVE_DUR,
ease: MOVE_EASE,
onUpdate: () => blurNode.setAttribute("stdDeviation", `${blurProxy.v} ${blurProxy.v}`),
},
MOVE_START,
);
Staggered grid streak-in (cards assemble)
For grid-card-assemble: each card streaks into its slot from its own backward offset, staggered. Drive every card off the same ease/window with a per-index delay; derive the entrance offset and start time from the card's index (no Math.random). Each card is sharp the instant it lands in its slot.
gsap.utils.toArray(".grid-card").forEach((card, i) => {
const at = MOVE_START + i * CARD_STAGGER;
tl.fromTo(
card,
{ x: ENTER_FROM_X, opacity: 0 },
{ x: 0, opacity: 1, duration: MOVE_DUR, ease: MOVE_EASE },
at,
);
// + a per-card blur proxy tween at the same `at` (Path A), or per-card ghosts (Path B)
});
Hold-the-streak (whip emphasis on a single beat)
For a single kinetic phrase that "zooms past," keep the streak slightly visible a frame or two longer by easing the blur on a marginally slower curve than the position (e.g. position expo.out, blur power3.out) — the element arrives, then the last wisp of smear resolves. Use sparingly; the default is locked envelopes.
How to Choose Values
Motion
- MOVE_EASE — shared ease for position and blur/echo.
- Range:
expo.out(hardest snap),power4.out(hard slam, default),power3.out(firm but softer) - Effects: harder
out→ velocity more front-loaded → blur reads as a sharper streak that resolves later in the window - Constraints: must be an
out-family ease (velocity front-loaded). AninOutorinease puts peak speed mid/late and the blur-speed coupling breaks. Position and blur must use the SAME ease (except the deliberate Hold-the-streak variation).
- Range:
- MOVE_DUR — travel + blur-resolve duration.
- Range: 0.25–0.6 s
- Effects: shorter → more violent whip; longer → a glide, the streak loses punch
- Constraints: a streak is fast — over ~0.7 s it stops reading as velocity blur and looks like a focus pull
- MOVE_START — timeline position of the entrance.
- Constraints: leave ≥1 s of dwell after
MOVE_START + MOVE_DURbefore the composition ends (climax dwell — a streak that lands att = DURATION − 0.2 sreads as "flashed and gone")
- Constraints: leave ≥1 s of dwell after
- ENTER_FROM_X / ENTER_FROM_Y — off-frame start offset along the motion axis.
- Range: 40–120% of the element's own dimension on that axis (far enough to read as "came from off-frame")
- Effects: larger → longer travel → the streak has more runway to read; too small and there's no sense of speed
Path A — SVG blur
- PEAK_BLUR —
stdDeviationat maximum velocity (start of the window).- Range: 8 (subtle) → 18 (default) → 30 (extreme whip)
- Effects: higher → heavier smear at peak speed; too high erases the glyph entirely at the start frame
- Constraints: ≤ ~30 — beyond that the element is unreadable for the first several frames and reads as "missing then appearing"; the filter region (
x/y/width/heighton<filter>) must be large enough (≥-50% … 200%) or the smear clips at the box edge
- SCALE_FROM (push-through variation) — starting scale for a camera push.
- Range: 1.3 (gentle push) → 2.5 (aggressive punch-through)
Path B — echo trail
- N (ghost count) — number of ghosts behind the lead (set by how many
.streak-ghostyou author).- Range: 2–4
- Effects: more ghosts → a longer, smoother smear; >4 reads as a stutter / strobe rather than a streak
- ECHO_STEP_PX — backward offset per ghost along the motion vector.
- Range: 12–40 px
- Effects: larger → a more spread-out, visible trail; smaller → a tight blur-like cluster
- Constraints:
(N) × ECHO_STEP_PXshould be ≲ENTER_FROM_Xso the furthest ghost still starts within the travel runway
- GHOST_BASE_OPACITY — opacity of the nearest ghost (
i = 1); falls off asBASE / i.- Range: 0.3 (faint) → 0.6 (pronounced)
- Constraints: ≤ ~0.6 — opaque ghosts read as duplicate elements, not a trail
Layout & type
- EL_FONT_SIZE / EL_TRACKING — the streaking element's type weight (when it's a phrase).
- Constraints: heavy display weight (≥120 px at 1080p, ≥800 weight) so the smear has mass to streak; thin type smears into invisibility
- CARD_STAGGER (grid variation) — delay between consecutive cards.
- Range: 0.05–0.12 s — tight enough to read as one assembling wave, not separate arrivals
Tokens
- {sceneBg} — background; a streak reads best against a solid / low-detail field (a busy bg fights the smear)
- {font} — typographic stack (embedded display face if the streaking element is text — see typography reference)
- {textColor} — element color; for Path B the ghosts inherit this, so a slightly desaturated trail can be had by tinting
.streak-ghostseparately - {phrase} — the word / glyph / wordmark that streaks in
Key Principles
- Blur peaks at peak speed, resolves to 0 at the settle — this is the whole rule. Share the ease and window between the position tween and the blur/echo envelope so they're locked. A blur that lingers after the element stops, or peaks after it's already slow, reads as a focus pull, not velocity.
out-family ease, always — velocity must be front-loaded (fast off the start, decelerating in).expo.out/power4.out/power3.out. AninorinOutease puts peak speed in the wrong place and the coupling falls apart.- Directional blur on the motion axis (Path A) —
stdDeviation="X 0"for horizontal,"0 Y"for vertical,"B B"only for a depth/scale push. A symmetric blur on a sideways move looks like defocus, not speed. - Tween a proxy, write the attribute (Path A) — GSAP tweens the plain
{ v }object;onUpdatecallssetAttribute("stdDeviation", …). You cannot tween the SVG attribute directly, and you must seed it once at setup so a seek tot=0shows the streaked start. - Ghosts are deterministic, by index (Path B) — offset
i * ECHO_STEP_PX, opacityBASE / i. NeverMath.randomfor the trail; index drives all per-ghost variation so every seek is identical. - Entrances only, never a mid-composition exit — a streak is an arrival. A blurred element leaving on a non-final frame reads as a glitch; scene-to-scene exits are the transition's job (see
../../transitions/overview.md). - Earn the sharp hold — after the snap, the crisp element must dwell ≥1 s. The contrast between the violent streak and the still, sharp settle is the effect.
- Heavy element, solid background — thin type or a busy backdrop both swallow the smear. Big bold mass on a clean field reads.
Critical Constraints
- Timeline must be paused:
gsap.timeline({ paused: true }). Nevertl.play(). - Registry key =
data-composition-idon the root. - No CSS
transitionon the streaking element (or ghosts) — it interpolates independently of HF seek and causes flicker. Only GSAP drives the move and the blur. - No
repeat/yoyo/ infinite — a streak is a single finite arrival. Finite tweens only. - No
Math.random/Date.now— ghost offsets/opacities and any stagger derive from the element index; deterministic every seek. - GSAP transform aliases only:
x,y,scale,rotation. Never tweenwidth/height/left/top. Tweeningfilter(the proxy →stdDeviation) andopacityis seek-safe and fine. - Seed the SVG
stdDeviationat setup (Path A) — write it once before play so a seek to the first frame renders the streaked start, not a momentarily-sharp pre-frame. - Filter region must be generous (Path A) —
<filter x="-50%" y="-50%" width="200%" height="200%">so the smear doesn't clip at the element's box edge. overflow: hiddenon the scene — the smear / furthest ghost extends past the resting position during travel; contain it so it doesn't bleed outside the frame.
Combinations
- kinetic-beat-slam.md — use this streak as the entrance for one phrase in a beat sequence (the scale-slam beat is a motion-blur fly-in); reads its onset from the shared
BEATS[]array - center-outward-expansion.md — the grid streak-in is center-expansion with a velocity-blur envelope on each element's travel
- 3d-text-depth-layers.md — extruded depth on the phrase that streaks in (depth layers ride the lead's transform)
- scale-swap-transition.md — alternative for a SAME-footprint state swap (this rule is for a fast ARRIVAL from off-frame / depth, not a morph)
Pairs with HF skills
/hyperframes-animation—out-family easing, proxy-drivenonUpdateattribute tweens, and locked-envelope coordination (../../adapters/gsap-easing-and-stagger.md)/hyperframes-creative—references/typography.md(embedded display face for a text streak),references/video-composition.md(solid field behind the smear)/hyperframes-core— composition wiring, determinism (finite tweens, noMath.random)/hyperframes-cli—hyperframes lint/hyperframes check(checkcatches a missing#streak-blurnode or an unreferenced filter)