Skip to content
solomotion
MakeSound and lip sync

Make it hear

Make a shape pulse with the bass, fire a pop on every bounce, draw a waveform, or make a mouth follow a voice.

Add and shape audio

  1. Import an audio file, or a video with sound. It becomes an audio layer.
  2. Trim it like any clip. Set fades of 20 to 50 ms on cuts to avoid clicks, and 1 to 3 seconds for music.
  3. Add audio effects to its chain: eq, reverb, echo, compressor, saturator, filter, chorus, phaser, tremolo, ring mod, bitcrush or stereo. Order is the signal chain.
  4. Route each clip to a category: Dialogue, music, effects or ambience.

Ducking is a rule on the music category, so music dips under the voice without keyframes.

Pulse to the music

Every audio layer is analysed once. Level, frequency bands, hits, beat and tempo are all readable live. Trimming or retiming the track keeps everything in sync.

  1. Select the layer to drive, such as its scale or opacity.
  2. Add an expression that reads the sound by layer name and feature: sound("Music", "bass").
  3. Shape it: 1 + 0.25 * sound("Music", "bass").

Features include level, sub, bass, lowMid, mid, highMid, treble, brightness, onset and beat. The audio inspector's Analysis section shows tempo and hits. beat and onset are already pulses, so do not smooth them.

For a visualizer, add a waveform layer that targets the track.

Sound the motion

  1. Select a layer, such as a ball with a body.
  2. Add a voice to its sounds stack.
  3. Choose a pack: pop, mechanical, thock, soft, vintage, digital, marimba, kalimba, glass, pluck or wood.
  4. Choose a trigger: text units, the layer's own in or out animation, a cut transition, or a hit.
  5. Set volume near 0.3 and randomness between 0.4 and 0.6 so repeats do not sound stamped.

A hit trigger fires once per collision, louder the harder the contact. Give every body the same voice with notes on a scale and the scene plays one melody.

Make a mouth talk

  1. Pick a character whose mouth is a switch or a set of shapes. Select it and open the Voice section. Choose Add mouth if none is set.
  2. Type a line in the line field to make an offline Test Voice clip. Its mouth timing is exact.
  3. Or use your own recording. The mouth follows whichever voice plays, or the one you choose in Voice.
  4. Under Engine, leave Auto, or pick another. Set Feel and Energy to taste.
  5. Choose the bake button when you want the mouth as keys to edit.

The Lip sync behavior hears the voice through the clip's own trim, speed and loop. Move, trim or swap the voice and the mouth follows. A missing mouth shape falls back to its nearest.

When you know the words, give the script to the analysis. It is far more accurate than listening.

Three functions read a voice for other things: lipsync("Voice") returns the mouth shape, speaking("Voice") is 1 while words are said, and loudness("Voice") is 0 to 1 for a jaw, a nod or a glow.

Ask an agent

Agents use media_import, audio_analyze, sound_category, sounds_list and sounds_use for audio. For speech they use speech_bind, speech_analyze, speech_inspect, speech_fix and speech_bake. Collision sounds are set on a layer with layer_patch.

Text
Bind the fox's mouth to the narration clip, give every ball a marimba hit sound, and make the ring pulse with the bass.