🎙️

Voice & Music

Script → voice → BGM → sync — a 20s audio pipeline with reusable settings

Task

Complete a 20s VO pipeline: script → pause marks → one voice → (optional) BGM → caption sync → rights check

Time

~22 min

Suggested tool: Hailuo AI

Self-rate (1–5)

How this scenario works

Audio fails usually mean scripts too long, inconsistent voice across clips, or BGM drowning VO. Fixed pipeline: 20s VO script (hook + selling point) → pause/pronunciation marks → pick one voice and preview → separate non-competing BGM draft → align captions in editor → commercial rights check. Good for: short-video narration, podcast intros, product VO drafts. Bad for: unauthorized voice cloning or impersonation; don’t put customer PII in scripts; confirm voice/BGM licenses before commercial use. Example: topic=urban coffee brand new drink; tone=relaxed and trustworthy; length=20s; need=VO + light BGM. Acceptable: 60–80 char script + pause notes; 3 voice directions; one 20s BGM prompt; alignment checklist. Tool pick: ElevenLabs for quality/multilingual; Descript for podcast-style editing; Hailuo AI for easier China access. Verify pricing on official sites.

Practice steps (with checks)

1Set acceptance: ~20s length, tone, BGM yes/no, commercial vs practice. Done when you know what “audio draft finished” means.
2Run “VO script”: 60–80 chars, 3s hook + core sell + light CTA. Done when punctuation is clear and lines aren’t tongue-twisters.
3Run “Pauses & pronunciation” for proper nouns and comma pauses. Done when a reader won’t guess pronunciation.
4Pick one voice in ElevenLabs/Descript/Hailuo and preview the full script — don’t switch mid-piece. Done when the whole track sounds consistent.
5If BGM is needed, run “BGM draft” as a separate non-competing 20s prompt. Done when BPM/mood/instruments are explicit.
6Align VO and captions in your editor using the alignment checklist. Done when subtitles are typo-free and sync looks OK.
7Before commercial use, run “Rights & ethics check” for voice/BGM. Done when you know publish vs practice-only.
8Save script + voice settings + BGM prompt; next time only swap [topic]/[tone]. Done when similar VO starts in ≤10 minutes.

Practice Workbench

Product / topicToneVO scriptUse caseBGM source

You edit short-video VO. Write ~20s Chinese narration for urban coffee brand spring launch, 60–80 characters.
Tone: relaxed, trustworthy, not hyped

Hook in 3s; clear punctuation; pronunciation notes for proper nouns; no invented product facts — use [TBD].

Output: [Title] [Script] [Tone] [Pause notes] [3 voice directions]

Open ElevenLabs

Outcome checklist

Self rating (1–5)

More prompts (5)

Pauses & pronunciation

Make this script easier to read aloud without changing facts:

[script]
Tone: [tone]

Output:
1) Marked script (/ short pause, // long pause)
2) Pronunciation table for proper nouns
3) Two awkward-line rewrites (same meaning, more spoken)

BGM draft

From this 20s VO, write a **non-competing** BGM prompt (ElevenLabs/Descript/Hailuo or general music models).

VO:
[script]
Mood: [tone]; ~20s

Include BPM range, lead instruments, mood, fade in/out; no vocal lyrics.
Output: Chinese prompt + English prompt + rough BGM level vs VO in the edit

Voice settings

You advise video sound design. Give copy-paste settings for ElevenLabs, Descript, or Hailuo AI.

Script:
[script]
Tone: [tone]

Output: 3 voice directions (M/F/neutral + emotion), speed hint, stability vs expressiveness, artifacts to avoid
Do not recommend products outside the listed tools.

Alignment checklist

I’m syncing VO and captions to picture. From this script, give a 6-item alignment checklist:

Script:
[script]

Items: first line within 3s, key caption highlights, BGM not drowning VO, end CTA readable, typo scan, export format hints (sample rate/mono suggestion)

Rights & ethics check

Review my audio plan for ethics/rights red flags (audit only):

Use: [use_case]
Script:
[script]
Plan: clone/library/custom recording; BGM source: [bgm_source]

List: risks, what to confirm before commercial use, safer alternatives (e.g., library voice vs clone). Do not invent legal conclusions.

Recommended tools

Best fitFreemium

ElevenLabs

Leading AI voice synthesis platform supporting multilingual text-to-speech and voice cloning.

AlternativeFreemium

Descript

AI-powered audio/video editor that lets you edit podcasts and videos like documents, with auto-transcription, filler removal, and AI voiceover.

Free-friendlyFreemium

Hailuo AI

MiniMax's AI voice and creation platform, supporting voice synthesis, AI music, and intelligent conversation.

After you leave with templates,

ChooseNarrow tools with Choose LabBrowse guides & compares in Lab Related learning pathCreate Images & Short Videos Try another scenario →

Continue your learning path

Practice a scenario Choose Lab

How this scenario works

Practice steps (with checks)

1Set acceptance: ~20s length, tone, BGM yes/no, commercial vs practice. Done when you know what “audio draft finished” means.

2Run “VO script”: 60–80 chars, 3s hook + core sell + light CTA. Done when punctuation is clear and lines aren’t tongue-twisters.

3Run “Pauses & pronunciation” for proper nouns and comma pauses. Done when a reader won’t guess pronunciation.

4Pick one voice in ElevenLabs/Descript/Hailuo and preview the full script — don’t switch mid-piece. Done when the whole track sounds consistent.

5If BGM is needed, run “BGM draft” as a separate non-competing 20s prompt. Done when BPM/mood/instruments are explicit.

6Align VO and captions in your editor using the alignment checklist. Done when subtitles are typo-free and sync looks OK.

7Before commercial use, run “Rights & ethics check” for voice/BGM. Done when you know publish vs practice-only.

8Save script + voice settings + BGM prompt; next time only swap [topic]/[tone]. Done when similar VO starts in ≤10 minutes.

Practice Workbench

Product / topicToneVO scriptUse caseBGM source

You edit short-video VO. Write ~20s Chinese narration for urban coffee brand spring launch, 60–80 characters. Tone: relaxed, trustworthy, not hyped Hook in 3s; clear punctuation; pronunciation notes for proper nouns; no invented product facts — use [TBD]. Output: [Title] [Script] [Tone] [Pause notes] [3 voice directions]

Outcome checklist

Self rating (1–5)

Make this script easier to read aloud without changing facts: [script] Tone: [tone] Output: 1) Marked script (/ short pause, // long pause) 2) Pronunciation table for proper nouns 3) Two awkward-line rewrites (same meaning, more spoken)

From this 20s VO, write a **non-competing** BGM prompt (ElevenLabs/Descript/Hailuo or general music models). VO: [script] Mood: [tone]; ~20s Include BPM range, lead instruments, mood, fade in/out; no vocal lyrics. Output: Chinese prompt + English prompt + rough BGM level vs VO in the edit

You advise video sound design. Give copy-paste settings for ElevenLabs, Descript, or Hailuo AI. Script: [script] Tone: [tone] Output: 3 voice directions (M/F/neutral + emotion), speed hint, stability vs expressiveness, artifacts to avoid Do not recommend products outside the listed tools.

I’m syncing VO and captions to picture. From this script, give a 6-item alignment checklist: Script: [script] Items: first line within 3s, key caption highlights, BGM not drowning VO, end CTA readable, typo scan, export format hints (sample rate/mono suggestion)

Review my audio plan for ethics/rights red flags (audit only): Use: [use_case] Script: [script] Plan: clone/library/custom recording; BGM source: [bgm_source] List: risks, what to confirm before commercial use, safer alternatives (e.g., library voice vs clone). Do not invent legal conclusions.