Sounds and music

flowchart LR
  A[Generate or import a clip] --> B[Review and save to library]
  B --> C[Place a saved variant]
  C --> D[Trim, fades and gain]
  D --> E[Check mix and export]

Playground and library

In Sounds & Music, choose an effect or music model, describe the sound and set its duration. Models may offer seed, lyrics or instrumental settings. Generate candidates, listen and save the useful version with a name and tags. You can also import an audio clip.

A generated candidate and a saved library sound are different records. Saving makes a reusable clip available to projects. Keep the settings with it when you want to create another variant.

Project cues

Add a saved sound to a project's audio. A cue can be inserted between speech or overlaid at a position. Trim it, adjust gain and apply fades where available. Listen across the cue boundary and the following line.

A project sound cue does not have its own ducking control. Voiceover ducking belongs to the source-track mix at export. Check the final assembled output, since individual previews do not show the whole balance.

Models and output

The sound catalogue lists model variants, backends and declared sound or music roles. A music-capable model may require lyrics or a supported instrumental mode; read its controls before a longer run.

Insertion and overlay

Choose insertion for an effect that needs its own space between spoken passages. It increases the output duration and moves later audio. Choose overlay for ambience or a music bed that should play under an existing passage; keep it within the output timeline.

Trim the useful region before placing a long clip. Start with a modest gain and short fades, then listen to the surrounding speech. A good library preview does not establish a suitable level in the project's final mix.

Reusing and changing a sound

A project placement refers to a saved library variant. Save a useful candidate before adding it to the project, and select the intended variant when a library item has several. Changing placement, trim or gain produces a new project mix; it does not require regenerating the speech.

Through MCP, the main tools are pandrator_sound_catalogue, pandrator_generate_sound, pandrator_sound_library and pandrator_project_sounds. Use pandrator_import_audio_reference for a local clip within an approved source root. Import accepts WAV, MP3, FLAC, OGG or M4A up to 16 MiB. These tool and import limits refer to release 0.12.0.

See sound effects and music.