prompt_h2501

Teaching Notes — Day 02 (Audio & Music Generation)

These notes support Day 02: Prompting for Music & Sound Generation. They include warm-up activities, prompt examples, and teaching cues for introducing audio generation tools and the PromptVision contest.

Block 1 — Foundations & Demos

Key Concepts Recap

• How AI generates audio: brief explanation of diffusion, autoregressive, and transformer-based models (e.g., MusicLM, Suno.ai, Udio).
• Modalities: text-to-music, text-to-sound, and audio-to-audio transformations.
• Prompt anatomy: genre → mood → tempo → instrumentation → structure → vocal style.

Talking Points:

• AI music models operate like text-to-image systems but in the time–frequency domain.
• Emphasize temporal composition: describe progression or mood change over time.
• Highlight ethical considerations (voice cloning, sampling, licensing).

Quick Demo Ideas

1. Genre variation demo:

   • Prompt: A calm ambient soundscape with ocean waves and slow synth chords.
   • Modify: A dark techno beat with ocean samples and metallic percussion. → compare rhythmic density and tone.

2. Mood contrast:

   • Prompt A: An uplifting orchestral theme with brass fanfare and fast tempo.
   • Prompt B: A melancholic piano piece with sparse strings and slow tempo.
   • Discuss: how prompt words like uplifting, melancholic, fast, slow shape harmonic and rhythmic character.

3. Instrument specificity:

   • A jazz trio with upright bass, brushed drums, and muted trumpet improvisation.
   • A synthetic pop track with electronic bass and vocoder vocals.

Optional Warm-Up Exercise

• Each student writes one prompt describing a 10-second soundscape.
• Instructor plays 3–4 outputs and asks the class: Does the audio match the prompt’s tone and instrumentation?

Block 2 — Tools & Prompt Craft

Key Platforms to Showcase

Type	Tool	Strengths
Text-to-Music	Suno.ai / Udio	High-quality multi-style songs with vocals.
Sound Design	Mubert / AudioCraft / AudioLDM	Great for ambient textures, SFX, and background loops.
Composition Aid	Soundful / Beatoven.ai	Quick structure-based generation, easy iteration.

Demonstration Flow:

1. Show one identical prompt across two models (e.g., Udio vs. Suno) → compare fidelity and style.
2. Discuss the difference between composition-level vs. texture-level control.

Example Prompt Templates

• A futuristic synthwave track at 110 BPM with arpeggiated bass and lush reverb.
• A cinematic trailer with strings, heavy percussion, and rising intensity.
• Lo-fi chillhop beat with vinyl crackle, warm piano chords, and laid-back tempo.
• Dark ambient drone with mechanical sounds and deep sub-bass.

Teaching Tip: Encourage iteration — changing one musical variable (tempo, mood, instrument) each time to observe the effect.

Block 3 — 🎤 PromptVision: AI Song Contest

Overview

Students participate in a Eurovision-style contest to apply prompting skills collaboratively.

Structure:

• 6–7 teams (same group size as Day 01).
• Theme announcement at start (e.g., Rebirth, Synthetic Love, Future Folk).
• Each team generates one 30–60s track matching the theme.
• All tracks are played in sequence; teams vote on creativity and prompt design.

Phases:

1. Theme reveal & brainstorming – 10 min
2. Prompt crafting & generation – 40 min
3. Submission & upload – 10 min
4. Listening party & voting – 40 min
5. Debrief – 10 min

Voting Categories:

• 🏆 Best Song (overall quality)
• 🎧 Best Prompt Concept (creative and clear)
• 🎭 Most Unexpected Genre Mix

Teaching Tips

• Limit track length to keep the session on schedule.
• Encourage teams to articulate why they chose specific prompt terms.
• For reflection, collect prompts privately (same rule as Day 01).
• Use a shared folder or Padlet board for track playback.

Example Prompts for Contest Inspiration

• A futuristic ballad about machines falling in love, 90 BPM, synthpop with vocoder vocals.
• A folk melody with AI choir singing in a fictional language, acoustic instruments, reverent mood.
• A chaotic drum & bass track symbolizing digital rebirth, 160 BPM, distorted samples.

Outcome for Day 02

After this session, students should be able to:

• Design prompts that control genre, mood, and tempo.
• Compare results across different music generation tools.
• Understand prompt iteration as compositional refinement.
• Critically evaluate creative and ethical aspects of AI-generated music.

This site is open source. Improve this page.