render-editorial-motion-podcast

Name: render-editorial-motion-podcast
Author: gooseworks-ai

1.0kstars193forks

How to install render-editorial-motion-podcast?

gooseworks-ai/goose-skills/render-editorial-motion-podcast

$npx -y skills add gooseworks-ai/goose-skills --skill render-editorial-motion-podcast

Installs into the current project.

Summary

Assemble an editorial-motion podcast-clip ad from a config — a real clipped podcast MP3 carries the narrative while N flat 2-tone editorial-illustration keyframes are animated NOT by generative i2v but by DETERMINISTIC ffmpeg ken-burns (zoompan) + hard cuts (no crossfades, which

Files · 1

View on GitHub

SKILL.md

1	# render-editorial-motion-podcast
2
3	Assemble an editorial-motion podcast-clip ad from a config: a real clipped podcast audio
4	line carries the whole narrative and every visual beat is timed to the sentence it describes,
5	in a bold flat 2-tone editorial-illustration look ("a New Yorker spot-illustration that
6	moves"). The motion is not generative video but deterministic ffmpeg ken-burns on static
7	keyframes, so it reads as a printed page that moves. This capability is that **FREE,
8	deterministic assembly** — the ffmpeg motion, hard-concat, audio mux, caption burn, and PIL
9	end card.
10
11	`scripts/config.example.json` is the worked example (Klarify "Rat Park", ~40.8s 1080×1920
12	9:16, 6 beats); `scripts/PIPELINE.md` maps every config block to its source step and
13	`scripts/README.md` documents the free assembly.
14
15	## Run
16
17	This is the FREE, deterministic assembly stage — it spends nothing on the motion layer.
18	The paid inputs are separate: the real podcast MP3 is clipped from source (free ffmpeg) with
19	its Whisper word timings, and one editorial-illustration keyframe per beat (chained ref images
20	so cage/character geometry holds) comes from `create-image-fal` (Nano Banana). Given the
21	clipped audio + `words.json` + the per-beat keyframes + the real brand wordmark PNG,
22	`render-editorial-motion-podcast` renders each keyframe as a ken-burns segment, hard-concats
23	on the beat, muxes the real audio, burns the mid-sentence captions, and composites the PIL end
24	card → the master. Re-cuts reuse the existing audio / keyframes and cost $0.
25
26	## Contract (the free assembly)
27
28	- A spoken narration carries the whole spot — no generated SONG. Mux the provided
29	narration MP3 (`-map 0:v:0 -map 1:a:0`) — a real clipped podcast line (preferred) OR an
30	approved generated VO (`create-vo-elevenlabs`). Never a sung/generated track. (Clip-vs-generate
31	is the recipe's STEP-0 intake decision — if no source episode is supplied, ASK the user.)
32	- NO generative i2v — deterministic ffmpeg ken-burns only. Animate each static keyframe
33	with `zoompan` (push-in / pull-back, 1.0→~1.06×, 24fps); Seedance/Kling are photoreal-trained
34	and invent naturalistic middle states that collapse the 2-tone look. Never `-loop 1` with
35	`zoompan d=N` (it balloons the duration); feed a single image and clamp with `-t` + `trim`.
36	- Hard cuts on the beat — no crossfades. Crossfades ghost two drifting cages through each
37	other; hard-concat each beat's segments and split long beats into micro-cuts (target 8–10
38	distinct visual moments). Each beat's visual STARTS within ~0.5s of its spoken line.
39	- Captions from Whisper word-timestamps, ON only mid-sentence. Burn `frosted-subtle`
40	captions while the speaker talks; leave silent/reflective beats and the end card uncaptioned.
41	THREE mandatory rules (each bit us in prod — bake them in):
42	1. NON-OVERLAP — clamp every line to END before the next STARTS
43	(`end = min(last_word_end + ~0.15, next_start - 0.03)`). Two boxes must never stack at the
44	same spot; an end-tail bleeding into the next window is the #1 caption bug.
45	2. SAFE AREA — captions sit in the lower third, so the keyframe's subject must stay in the
46	upper ~75% (see the recipe's `look_pack.caption_safe_area`). If a finished keyframe's subject
47	intrudes into the caption band, deterministically shift the subject UP into the empty top
48	space (PIL: paste up ~0.24H onto a canvas pre-filled with the exact paper color from a clean
49	corner) — never let the box sit on the subject.
50	3. BURN ENGINE — prefer libass (`ass`/`subtitles` filter), but **check `ffmpeg -filters`
51	first**: many builds (Homebrew) lack libass/drawtext. If absent, use the deterministic
52	overlay fallback — render each line as a transparent PNG (frosted rounded box + white
53	text, PIL) and composite via the ffmpeg `overlay` filter with timed
54	`enable='between(t,st,en)'` windows. Same look, no libass.
55	- End card via PIL from the real wordmark PNG — never AI-render brand text. The lockup is
56	composited deterministically (stretched-gradient bg + feathered mascot crop + wordmark +
57	tagline with a system font); a diffusion model garbles a wordmark

People who liked this also viewed

6 picks

Type

render-editorial-motion-podcast

1.0kstars193forks

View on GitHub

How to install render-editorial-motion-podcast?

gooseworks-ai/goose-skills/render-editorial-motion-podcast

$npx -y skills add gooseworks-ai/goose-skills --skill render-editorial-motion-podcast

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# render-editorial-motion-podcast
2
3	Assemble an editorial-motion podcast-clip ad from a config: a real clipped podcast audio
4	line carries the whole narrative and every visual beat is timed to the sentence it describes,
5	in a bold flat 2-tone editorial-illustration look ("a New Yorker spot-illustration that
6	moves"). The motion is not generative video but deterministic ffmpeg ken-burns on static
7	keyframes, so it reads as a printed page that moves. This capability is that **FREE,
8	deterministic assembly** — the ffmpeg motion, hard-concat, audio mux, caption burn, and PIL
9	end card.
10
11	`scripts/config.example.json` is the worked example (Klarify "Rat Park", ~40.8s 1080×1920
12	9:16, 6 beats); `scripts/PIPELINE.md` maps every config block to its source step and
13	`scripts/README.md` documents the free assembly.
14
15	## Run
16
17	This is the FREE, deterministic assembly stage — it spends nothing on the motion layer.
18	The paid inputs are separate: the real podcast MP3 is clipped from source (free ffmpeg) with
19	its Whisper word timings, and one editorial-illustration keyframe per beat (chained ref images
20	so cage/character geometry holds) comes from `create-image-fal` (Nano Banana). Given the
21	clipped audio + `words.json` + the per-beat keyframes + the real brand wordmark PNG,
22	`render-editorial-motion-podcast` renders each keyframe as a ken-burns segment, hard-concats
23	on the beat, muxes the real audio, burns the mid-sentence captions, and composites the PIL end
24	card → the master. Re-cuts reuse the existing audio / keyframes and cost $0.
25
26	## Contract (the free assembly)
27
28	- A spoken narration carries the whole spot — no generated SONG. Mux the provided
29	narration MP3 (`-map 0:v:0 -map 1:a:0`) — a real clipped podcast line (preferred) OR an
30	approved generated VO (`create-vo-elevenlabs`). Never a sung/generated track. (Clip-vs-generate
31	is the recipe's STEP-0 intake decision — if no source episode is supplied, ASK the user.)
32	- NO generative i2v — deterministic ffmpeg ken-burns only. Animate each static keyframe
33	with `zoompan` (push-in / pull-back, 1.0→~1.06×, 24fps); Seedance/Kling are photoreal-trained
34	and invent naturalistic middle states that collapse the 2-tone look. Never `-loop 1` with
35	`zoompan d=N` (it balloons the duration); feed a single image and clamp with `-t` + `trim`.
36	- Hard cuts on the beat — no crossfades. Crossfades ghost two drifting cages through each
37	other; hard-concat each beat's segments and split long beats into micro-cuts (target 8–10
38	distinct visual moments). Each beat's visual STARTS within ~0.5s of its spoken line.
39	- Captions from Whisper word-timestamps, ON only mid-sentence. Burn `frosted-subtle`
40	captions while the speaker talks; leave silent/reflective beats and the end card uncaptioned.
41	THREE mandatory rules (each bit us in prod — bake them in):
42	1. NON-OVERLAP — clamp every line to END before the next STARTS
43	(`end = min(last_word_end + ~0.15, next_start - 0.03)`). Two boxes must never stack at the
44	same spot; an end-tail bleeding into the next window is the #1 caption bug.
45	2. SAFE AREA — captions sit in the lower third, so the keyframe's subject must stay in the
46	upper ~75% (see the recipe's `look_pack.caption_safe_area`). If a finished keyframe's subject
47	intrudes into the caption band, deterministically shift the subject UP into the empty top
48	space (PIL: paste up ~0.24H onto a canvas pre-filled with the exact paper color from a clean
49	corner) — never let the box sit on the subject.
50	3. BURN ENGINE — prefer libass (`ass`/`subtitles` filter), but **check `ffmpeg -filters`
51	first**: many builds (Homebrew) lack libass/drawtext. If absent, use the deterministic
52	overlay fallback — render each line as a transparent PNG (frosted rounded box + white
53	text, PIL) and composite via the ffmpeg `overlay` filter with timed
54	`enable='between(t,st,en)'` windows. Same look, no libass.
55	- End card via PIL from the real wordmark PNG — never AI-render brand text. The lockup is
56	composited deterministically (stretched-gradient bg + feathered mascot crop + wordmark +
57	tagline with a system font); a diffusion model garbles a wordmark

People who liked this also viewed

6 picks

Type