Voronova · NASA Space Apps story reel · vertical 1080x1920

Notes for the editor

Each section starts with reference frames to match, with a short note under each, followed by the details. Mock frames are rough guides made from the existing video, so recreate them properly from the original photos. Every voice line is below as its own file, so you can move or trim them freely.

Start here

Current rough cut (2:23)

720p preview. Download · Full voice track only: voronova_vo_full.mp3

Beat 0 · "So it was 7pm, I was lying in bed writing code, when my friends from Punjab University called" · rough cut ≈ 0:08–0:13

Replace the bed shot with a better clip

REPLACE THISThe current shot of Waleed lying in bed (original 0:07–0:11).
  • Swap it for a better clip of the same moment from Waleed's Google Drive. Waleed will share the folder link. Pick the sharpest, best-lit take, ideally with the laptop and code visible, and with real motion rather than a still.
  • Keep the existing "it was 7 PM / I was lying" text animation on top, and match the warm night grade of the surrounding shots.
  • Cut to the phone call on "called" (the Punjab University friends photo follows).

Beat 1 · right after "…just kidding, Sharafat" · rough cut ≈ 0:44

Team intro, movie-style (with motion)

1 · STARTFull team photo, as it sits in the video now.
2 · PUSH INSmooth camera zoom into Ahad (left, striped shirt). Ease-in-out, slight drift.
3 · MOVIE LOOKLetterbox bars. Rest of frame dark and grey, Ahad in colour with a white cut-out outline. Title AHAD / "the neighbor".
4 · MOVE TO SHARAFATWhip-pan or glide across, same treatment. Title SHARAFAT / "the one who said i'm in".
5 · PULL BACKBars and grade come off, back out to all three on "perfect pick".
"Ahad. just a neighbor who was down for it. and Sharafat, another dude who just said i'm in. both could design quite well tho, so i thought they'd be the perfect pick."
  • The photo must move the whole time. No static holds. Keep the camera slowly pushing even while a title is up.
  • Outline: rotoscope each person (or use an AI cut-out) and add a 6–8 px white stroke. Animate the stroke drawing on over about 8 frames.
  • Grade: background desaturated and about 65% darker, subtle film grain, slightly warm highlights on the person.
  • Titles: high-contrast serif (Didot or Bodoni style), large and tracked out a little, with a small subtitle under it. Fade/slide up in about 10 frames, timed to the name in the VO, with a soft whoosh or low hit on each name.
  • Waleed (right, black t-shirt) stays dimmed for both intros.
  • The colour frame in the video has "Straight to campus" text over it, and the mock has a smudge where we removed it. Use the original photo (ask Waleed).

Beat 2 · "So I called in the big guns, Jensen Huang… just kidding, Sharafat"

Jensen Huang joke

MATCH THISJensen on stage in the leather jacket, wide or medium shot. GTC 2026 highlights. youtube.com/watch?v=kDd24YOeqQQ
FULL KEYNOTE2 h 18 m official upload, for more angles. Download at the highest resolution available. youtube.com/watch?v=jw_o0xr8MWU
  • On "Jensen Huang": about a 1 s cut of Jensen on stage, ideally him turning or gesturing to camera.
  • On "just kidding": record scratch or freeze plus a quick desaturate, then hard cut to Sharafat's photo.
  • Keynote page: nvidia.com/en-us/gtc/keynote

Beat 3 · "we picked the Home in Space challenge. we wanted to save the time it takes to make 2D and 3D assets for designing space habitats" · rough cut ≈ 0:58

Home in Space: habitat b-roll (NASA)

MATCH THISHabitat interior greenhouse corridor. Best fit for "colonies for scientists". NASA "Living and Working on Mars", around 2:40. direct mp4, 720p
ALTAstronaut and rover on Mars, same video, around 1:30.
ALTMars-from-orbit opener, same video, around 0:30. page

Beat 4 · "so we made a conversational agent, fine tuned on NASA's drawings and models. it generates an editable 2D habitat layout, and once the design is final, it spits out a 3D model" · rough cut ≈ 1:03–1:11

Our AI: 2D layout → 3D model (TRELLIS)

Show the flow: chat prompt → 2D floor-plan layout being edited → 3D model building up. Use the TRELLIS footage for the 3D part. For the chat and 2D layout, a quick screen recording of our app works best (ask Waleed).

MATCH THISModel building up layer by layer ("internal structure"). Use for "turns it into a 3D habitat". TRELLIS.2 demo, around 0:40.
ALTWall of generated assets: "it can make anything". Around 0:25.
ALTMaterial variants. Around 0:12.
ALTClose-up renders. Around 0:52.

Beat 5 · after "No internet, even Canva stopped working" · original 0:52

"And this was us" (dino game)

SOURCEExisting laptop shot. The offline dino game is top-middle.
MATCH THISFull screen, punched into the dino and cactus. No text and no labels.
"and this was us."
  • Show only that clip, full screen, with a slow push-in. Viewers work out the game from the footage. Don't add captions explaining it.

Ending · from the winners announcement to the last line

Ending order

  1. "Then they announced the winners" → "Honestly, we never thought we'd win…" over the crowd.
  2. "Team Voronova" on the team walking up.
  3. "We did it" goes right before the speech, the clip of the team at the front with the mic (original ≈ 1:38–1:46). Let the speech play with its own audio (music ducked) or a short bit of it.
  4. After the speech: "So here's the point. If a 15 year old can pull this off, you can too. Keep hope in you." This is the last line, over the end cards.

If the video runs too long

Cutting it down, or splitting into 2 parts

The current cut is 2:23. Under about 1:30 holds attention best on Reels and TikTok. Either trim (option A) or post as a 2-part series (option B). Times below are from the current rough cut and will shift a few seconds after the new lines go in, so find each line by its words.

Option A: trims, in order (do the first ones first)

  1. Music gap after "Team Voronova" (1:57–2:05): shorten 8 s to about 3 s. Keep the team walking up. Saves ~5 s.
  2. KFC line "Years of those late night KFC runs…" (1:13–1:17). The picture is a static hold here anyway. ~5 s.
  3. "At events like Google, Figma, Canva, you keep meeting the same kind of people" (0:13–0:18). ~4.5 s.
  4. "10x hackathon winners, cracked folks I'd never met in Pakistan": keep "19 challenges, people had been preparing way before the competition" and cut the second half (≈0:28–0:31). ~3.5 s.
  5. "Questions were supposed to be 5 minutes, it went on for 15" (1:36–1:39). ~3.5 s.
  6. Tail of "honestly…": end at "making jokes" and drop "talking with people, learning from their experiences, it was fun" (≈1:49–1:54). ~4 s.
  7. Einstein/Elon line (0:37–0:41): only if still long. The Jensen joke right after carries the same humour. ~4.5 s.

All seven together save about 30 s, bringing it to roughly 1:47. Cut on breaths or pauses, never mid-word, and let the music run under the cuts so they don't jump.

Never cut: the hook ("We won…, team lead…"), the team intro, Home in Space + the agent, "everything broke" / "and this was us", winners → Team Voronova → "we did it", or the closing line.

Option B: 2-part series

  • Part 1 (≈1:20): "the build". Hook + team lead → finding the hackathon → team intro (Ahad, Sharafat) → Home in Space + the agent → all-nighter → "Then straight to the campus, and of course everything broke." Stop on "everything broke" as the cliffhanger, then a 1 s end card: "Part 2" (same typography as the reel).
  • Part 2 (≈55 s): "the win". Opens with Waleed's line "part two. so we got to campus, and everything broke." (ready-made file: vo_part2_opener.wav) over the campus photo → "No internet, even Canva stopped working" → "and this was us" (dino) → presenting, jokes, questions, phones → winners → honestly → Team Voronova → "we did it" → closing line.
  • Each part keeps the music bed. Re-time the music so Part 1's build peaks on "everything broke" and Part 2's peaks on "Team Voronova".
  • Captions: Part 1 cover text "how a 15 year old won NASA Space Apps (part 1)". Part 2: "part 2: the presentation".
  • Apply the Option A trims inside each part too, if needed.

Already done in the voice track

Other changes

  • Removed "At least there were snacks and no parents complaining…".
  • Removed "so what can we do?" after "they were forced to listen anyways".
  • "Team Voronova" lands on the team walking up (original ≈ 1:34), followed by about 8 s of music only, then "We did it".
  • Music ducks under the voice. Final mix target is -14 LUFS.

Who's who: Ahad = left (grey striped shirt), Sharafat = middle (brown shirt, behind the banner), Waleed = right (black t-shirt).

All voice-overs

Voice lines (play, download, rearrange)

Each line from the final voice track is a separate file, numbered in story order, with the time it sits at in the rough cut. Below them are the raw recording takes, which include alternate deliveries and ad-libs, if you want a different read.

Extra

Raw takes (alternates)

Need a line changed or added?

Request a voice-over

Write the exact line you need and where it goes. Waleed records it and it shows up in the voice lines above.