speech_only preset

Extract only speech portions — removes music, singing, and silence.

What it keeps

• Speech segments (clear talking without music)
• Talking over music (speech overlaid on music)
• Speech over music (general speech+music mix)

Usage

praisonai-editor edit radio_show.mp3 \
  --preset speech_only \
  --detector ensemble \
  -v

When to use

• Radio shows where you want interview clips without music beds
• Live stream recordings — keep only the talking parts
• Music shows — extract the host commentary only

Note

Requires --detector (a content detector). Use ensemble for best results.

Python API

from praisonai_editor.pipeline import edit_media

result = edit_media(
    "radio.mp3",
    preset="speech_only",
    detector="ensemble",
    verbose=True,
)