Module 09 — Captions, Voice, Music, and Sound Design

Category: Finishing | Duration: ~75 min | Prerequisites: Module 08

Learning Objective

By the end of this module, you will be able to make videos accessible, intelligible, and emotionally engaging by adding accurate captions, balancing voice-over/music/SFX levels, selecting mood-appropriate music, adding reinforcing sound effects, using strategic silence, and verifying commercial-use licensing. You will produce two sound treatments for the same edit and compare them.

1. Principle: Audio Serves Comprehension First, Emotion Second

A video that can't be understood silently is a video that fails. A video with competing audio is a video that frustrates.

Caption-First Accessibility

Every video must be fully understandable without sound when the concept relies on spoken information.

Accurate Transcription

1. Listen to the full voice-over and transcribe every word
2. Do NOT auto-transcribe — AI transcription introduces errors
3. Correct every error manually
4. Verify speaker identification if there are multiple voices

Caption Readability

Caption Punctuation and Line Breaks

• Break on clauses, not words: You see an ambulance / go by. You never see this.
• Use commas within lines, periods between complete thoughts
• Break at natural pauses in the voice-over
• Use em dashes or colons for speaker changes (if applicable)

Voice-Over Standards

Voice-over must always be audible over music and effects. The 2:1 ratio (VO is 6dB louder than music) is the minimum acceptable.

Music Selection

1. Match the mood of the scene — not the editor's taste
2. Check the genre — a romance needs different music than a thriller
3. Avoid lyrics during voice-over (they compete for attention)
4. Use instrumental versions when lyrics might distract
5. Match the tempo to the pacing of the edit

Sound Effects

Sound effects should reinforce meaningful visual actions, not fill silence:

Avoid: Generic "whoosh" templates applied to every transition, music that overpowers the story, SFX that are louder than the VO.

Strategic Silence

Silence is a powerful tool. Use it:

• Before a reveal — builds tension
• After a payoff — lets the insight land
• Between beats — creates rhythm
• When the visual carries the story — no VO needed

Licensing Verification

Never assume that a track available within one platform is automatically cleared for use on every other platform.

Verification Protocol

For every audio asset:

1. Check the license type — commercial use? Redistribution? Platform restrictions?
2. Verify the source — royalty-free library? Custom recording? AI-generated?
3. Document the license — save the license file or URL
4. Confirm platform eligibility — can this be used on TikTok AND YouTube AND Instagram?
5. Check expiration — do licenses expire? (They do.)

Approved Sources

If narration is not part of the concept, ensure the visuals and on-screen text still communicate a complete idea. A silent video with no text is not accessible.

2. Visual Demonstration

Professional Example — Sound Treatment Comparison

Book: Untold Stories of a Paramedic Concept: The Overdose (naloxone administration)

Treatment A — Full Sound Design

Result: Every sound supports the story. VO is always audible. No competing elements.

Treatment B — Poor Sound Design

Result: VO inaudible, music overpowers, SFX distract. Silent viewing fails.

3. Step-by-Step Workflow

1. Transcribe the Voice-Over

1. Play the full voice-over track
2. Write down every word accurately
3. Note any mispronunciations (correct in post-production)

2. Add Captions

1. Import the transcription into the editing tool's caption track
2. Sync each caption to the corresponding voice-over line
3. Break lines at natural pauses (max 30 chars per line)
4. Set display duration (min 1.5 seconds)
5. Verify contrast against the background on every frame

3. Level the Voice-Over

1. Check the VO level (-12 to -6 LUFS)
2. Use compression to smooth dynamics
3. Apply noise reduction if needed
4. Do NOT over-compress (kills natural speech dynamics)

4. Add Music

1. Import the selected music track
2. Set to -18 to -12 LUFS
3. Use the audio ducking feature if available (music lowers when VO plays)
4. If no ducking: manually lower music during VO lines

5. Add Sound Effects

1. Place SFX at the exact moment they occur visually
2. Set to -20 to -15 LUFS
3. Use SFX that reinforce, not duplicate, the visual
4. Remove any SFX on silent beats (let silence breathe)

6. Verify Licensing

1. For every audio asset, confirm the license
2. Save license documentation in 02_Source-Metadata/
3. Confirm platform eligibility
4. Flag any unverified assets as "needs approval"

7. Test Playback

1. Listen on speakers — check balance
2. Listen on headphones — check spatial placement
3. Watch with sound off — verify captions carry the story
4. Check on mobile — verify caption size and contrast

4. Common Mistakes and Corrections

5. Hands-On Exercise

Produce Two Sound Treatments

Using the video you assembled in Module 08, create two distinct sound treatments:

Treatment A — "Full Production"

• Voice-over at -9 LUFS
• Music at -16 LUFS (instrumental, no lyrics)
• Sound effects on key actions
• Strategic silence where appropriate

Treatment B — "Minimalist"

• Voice-over at -9 LUFS
• No background music (sound effects only)
• Strategic silence as a design element

Comparison Analysis

For each treatment, analyze:

Licensing Check

For Treatment A's music, complete the licensing verification protocol:

1. Source: ________________
2. License type: ________________
3. Commercial use: Yes / No
4. TikTok eligible: Yes / No
5. Instagram eligible: Yes / No
6. YouTube eligible: Yes / No
7. License documentation saved: Yes / No

6. Assessment

Scored assessment (40 points total):

Pass threshold: 34/40

Submission Requirement: Submit both sound-treated versions of your video, the comparison analysis, and the licensing verification. The video must remain understandable without sound when the concept relies on spoken information.

7. Multiple Narration Voices Across a 15-Video Batch

A single default Google Vids voice must not narrate all 15 videos. Voice assignment is a production decision, not a promise that a particular number or type of voices is available in every account.

Required Voice Workflow

After importing the document and generating the initial material:

1. Identify the narration automatically created by Google Vids.
2. Replace or edit it into the approved voiceover for each scene.
3. Select suitable alternative voices from the options actually available.
4. Deliberately distribute voices across the batch.
5. Ensure the voice matches the book's audience, genre, and intended tone.
6. Check pronunciation of names, pacing, clarity, and synchronization.

Suggested Voice Allocation

Use five voice groups, with three videos per group, where the available voices and book content support that arrangement. Reuse voices intelligently when needed, but avoid one voice throughout the entire batch.

Voice Consistency and Pronunciation

• Keep the narration for each individual video internally consistent unless an intentional, documented change is needed.
• Document every voice assignment in the master brief and batch production template.
• If voice availability is limited, document the verified options and an approved fallback workflow.
• Check the actual rendered audio duration against the 10-second scene; adjust wording, pacing, and pauses as needed.

Downloadable Resources

• Caption Legibility Checklist
• Resource Library: Audio Licensing Guide
• Batch Production Template