A good ha is all about timing, clarity, and perspective. Use short, dry hits when the laugh is a punctuation mark, and longer runs when the laugh becomes part of the rhythm. When generating, describe the syllable count (single ha vs ha ha ha), intensity, and space so the transient lands on your cut and the tail does not step on the next line.
Choose the role: hit, run, or crowd
Start by deciding whether the laugh is a single accent, a short sequence, or a room reaction. This determines how much decay you can afford and how prominent the voice should feel in the mix.
- For a punchline button, ask for a single ha with fast decay and minimal room tone
- For pacing, request a laugh run with natural gaps between syllables
- For audience energy, specify a small crowd and a wider stereo image
Control tone and intelligibility
Ha can read as friendly, sarcastic, awkward, or villainous depending on pitch, brightness, and breath texture. Describe these traits so the sound fits the character and does not distract.
- Use bright, crisp delivery for quick meme tags and sharp edits
- Use low, warm delivery for grounded dialogue scenes
- Add breathy or snorty texture for awkward or goofy character beats
Match distance and space to picture
A close laugh with no reflections feels like it is in the viewer's ear, while a distant laugh needs softened transients and light room reflections. Ask for the perspective you need so it sits naturally under the scene.
- Close mic: narrow stereo, clean attack, almost no tail
- Across the room: softer transient, light reflections, short natural decay
- Crowd bed: layered voices, wider stereo, controlled room bloom
Avoid common ha mistakes
Overly long tails, noisy backgrounds, or unnatural cadence can make a laugh feel fake. Use prompt constraints that keep the clip easy to cut and believable.
- Avoid long reverb tails unless the scene clearly shows a large room
- Avoid heavy background noise that will fight music or dialogue
- Avoid robotic timing by asking for natural spacing and human breath
