1 Course position and tonight’s promise

Day 5 continues Module 2. Tonight you will turn a fictional wedding invitation into a 30-second video made of six five-second shots, add a hook, voiceover, captions, and music, then export it in the right ratio for WhatsApp status—checking credits and clip limits the whole way.

2 What you will learn and why it matters

• Plan an idea-to-export pipeline: script, voice, shots, timeline, captions, export.
• Open with a three-second hook and end with a clear call to action.
• Add a multilingual AI voiceover and check pronunciation before export.
• Write text-to-video shot prompts and animate a still photo with image-to-video.
• Use a talking avatar only with consent, and keep the credit meter in view.
• Export in the aspect ratio each platform actually needs.

A short video is a sequence of small, checkable decisions—not one big generate button. Each clip, voice line, and caption needs its own verification.

3 A familiar situation: diagnose before you render again

Fictional sample values: Arjun and Meera's wedding invitation needs a 30-second WhatsApp status video. The first cut has three visible problems. Predict the missing fix before revealing it; do not use a real couple's name, contact number, or venue address.

4 Simple explanation, then accurate vocabulary

A hook is the first few seconds that earn attention; without it, most viewers scroll past. A voiceover is a spoken audio track added to a video. Text-to-video generates a clip from a written description, while image-to-video animates motion from a still photo you already have.

5 Idea-to-export pipeline

A dependable short video follows a pipeline: idea → script → voice → shot list → clips → timeline → captions → export. Skipping a stage—like exporting without a shot list—makes the result hard to predict and expensive to redo, since regenerating clips can use limited credits.

6 Three-second hook and call to action

The first three seconds decide whether a viewer keeps watching. Open with the most visually interesting or emotionally clear shot, not a slow introduction. Close with a clear call to action: what should the viewer do—save the date, tap a link, or reply?

7 Multilingual AI voiceover

An AI voiceover can narrate in multiple languages, but names, places, and honorifics are the most common mispronunciations. Listen to the full line before export, check every proper noun against how it is actually said, and re-generate or manually adjust phonetics if needed.

8 Text-to-video shot prompts

A text-to-video shot prompt needs the same anatomy as an image prompt, plus timing and motion: subject, style, setting, camera movement, duration, and constraints. A vague prompt like "wedding scene" produces an unpredictable five seconds; a specific one is reviewable before it uses credits.

Shot: warm indoor mandap scene, soft golden lighting, slow pan left.
Duration: 5 seconds. Style: soft cinematic, no readable text.
Constraint: no real faces, no invented names or logos on screen.

9 Image-to-video

Image-to-video takes one still photo—such as the restored invitation photo from Day 4—and adds motion: a gentle zoom, pan, or parallax effect. It works best with a clean, high-resolution source image and a small, natural amount of movement; large motion requests often distort faces or backgrounds.

10 Talking avatars

A talking avatar is a generated on-screen presenter or character that can read a script aloud. Using a real person's likeness as an avatar needs their explicit consent, the same as a family-photo greeting. For a fictional or licensed character avatar, still verify the script and pronunciation before export.

• Get explicit consent before creating an avatar of a real person.
• State clearly if a video uses a generated avatar rather than the real person.
• Check the avatar's script line-by-line for accuracy, not just tone.
• Avoid an avatar for sensitive announcements where a real, verified voice matters more.

11 Editing, music, captions, venue card

The timeline is the ordered sequence of clips, audio, and text in the edit. Arrange the six shots in story order, add background music under the voiceover at a lower volume, generate captions for accessibility, and insert a clean venue/date card near the end so the key facts are visible, not just spoken.

12 Platform aspect ratios

Different destinations need different shapes: WhatsApp status and most short-form feeds expect a tall 9:16 frame, a shared status board or TV screen may need 16:9, and some chat previews favour a square 1:1. Choose the ratio before generating shots, because reframing afterward can crop faces or text cards.

13 Personalised birthday songs and name/pronunciation checks

A personalised song for a birthday or celebration video reuses this same pipeline, but names and nicknames are the highest-risk detail: a mispronounced name in a song is far more noticeable than in speech. Always request a spoken preview of the exact name before generating the full song, and confirm the preferred pronunciation with the family first.

• Confirm the preferred name/nickname and pronunciation with the family directly.
• Generate a short spoken preview of the name before the full song.
• Re-check age-appropriate lyrics for a child's video specifically.
• Keep the child's full name and school out of any publicly shared version.

14 Short-clip and credit limits

Most video tools limit clip length (often five to ten seconds per generation) and charge credits per generation, voice line, or export. Plan the full shot list and word count before generating, watch the credit meter as you go, and treat a small-business promotional video as a commercial use needing its own plan and rights check, separate from a personal invitation.

15 Interactive process model: script, shoot, assemble, verify

The idea card is the learner's real input; the script/shot-list is the planning decision; the clip-generation card is the tool's attempt at each five-second shot; the review gate checks pronunciation, motion, and consent; the timeline-assembly stage adds music and captions; the last card is the exported video. This simplified picture does not expose hidden model reasoning.

16 Facilitator-led live demonstration: one ratio mismatch

Start with the fictional 16:9 cut exported for WhatsApp status. Intentionally keep the wide ratio because it "looked fine in the editor preview." Show why that is a mismatch: a phone status frame is tall, not wide. Diagnose the black bars and regenerate the shot list and export settings for 9:16.

• Use the tool's current New project or Edit equivalent with a fictional shot list.
• Export the 16:9 draft and identify the black bars on a phone-sized preview.
• Rebuild the project with a 9:16 canvas and the same shots, voice, and captions.
• Compare the new export against an actual phone-status preview.
• Save the before/after and state why the ratio decision changed.

17 Your Video Invitation Production Sheet

Save this as “Day 5 — Video Invitation Production Sheet.” For each fictional item, keep the prompt, the verification evidence, and a one-line reflection.

18 Learner lab: build your own safe 30-second video

1. Prepare one fictional event's facts and, optionally, one non-sensitive practice photo.
2. Write a six-shot list with a three-second hook and a clear call to action.
3. Generate voiceover, text-to-video, or image-to-video clips one shot at a time.
4. Listen to every spoken name and inspect every motion effect before accepting a clip.
5. Assemble the timeline with music, proofread captions, and a checked venue/date card.
6. Export in the correct aspect ratio for your destination; add a one-line reflection.

19 Behind the scenes, limitations, safety, and verification

• Generated clips rarely match a prompt exactly; review each one before adding it to the timeline.
• Voiceover pronunciation is frequently wrong for names, places, and honorifics.
• Image-to-video can distort faces or objects, especially with large requested motion.
• A talking avatar of a real person is not that person's actual current words or consent.
• Clip length and export credits are limited and vary by tool and plan.
• Style, voice, and quality vary by tool and version; today's result may not repeat tomorrow.

20 Common confusion, errors, troubleshooting, and best practices

• Best practice: write the full shot list and script before generating a single clip.
• Best practice: check names, motion, and consent before assembling the timeline.
• Trainer tip: ask “What in this video was verified by me, and what was only generated?”

21 Takeaways

• Plan the full idea-to-export pipeline before generating a single clip.
• Open with a three-second hook and close with a clear call to action.
• Listen to every voiceover name and inspect every animated motion before trusting it.
• Get explicit consent for any real-person avatar or voice.
• Export in the aspect ratio the destination actually needs, and track credit usage.

Module 2 continues when you can explain what your shot list controlled, what you verified, and what rights or consent you checked.

22 What comes next

Next is Topic 06—Create Presentations and Documents. You will carry this module's verified facts and design habits into slide decks and documents, checking formatting, sourcing, and rights the same way before Module 2's assessment.

• Reopen your Video Invitation Production Sheet.
• Keep your verified event facts and chosen style ready.
• Note one formatting or sourcing question you predict the next lesson will need to answer.