COMING SOON · DEVELOPMENT PREVIEW
Podcast AI editing for Adobe Premiere Pro
A speaker-driven camera planning workflow is in development for interviews and 2–3 person podcasts. It is not available in current public downloads.
By Rizki Ardiansyah, creator of R Project Multicam AI · Updated 8 October 2026
What is working in the local test build?
- Decode isolated microphone audio locally with FFmpeg.
- Use microphone activity to plan switches between mapped speaker cameras.
- Set a minimum shot duration to reduce rapid camera changes.
- Use an optional wide camera for overlap and sustained silence.
- Preview the plan before applying cuts and disabling unused angles with Keep Clips.
The planner and FFmpeg decoding passed synthetic-audio tests. Real-recording quality and actual Premiere timeline execution have not been validated. These tests do not establish production readiness.
How the intended podcast workflow differs from music editing
Music multicam uses beat timing, BPM and drop candidates. Podcast planning instead uses microphone activity to choose a speaker camera. They solve different editing problems; a music beat map does not identify who is speaking.
Prepare footage for the test workflow
- Synchronize cameras and isolated microphone tracks before analysis.
- Give each speaker a separate microphone audio track and map it to a camera.
- Work in a saved, duplicated sequence.
- Generate and inspect a short camera-plan preview.
- Only apply to a test sequence, then review speaker timing and camera coverage.
Current limitations
This is microphone-activity analysis, not semantic understanding, transcription or facial recognition. It does not identify separate speakers from a single mixed microphone track. Microphone bleed, music and unequal recording levels can lead to incorrect switches.
Video-based active-speaker detection is a possible future research direction, not an implemented feature. No release date, operating-system readiness or performance guarantee has been announced.
Frequently asked questions
Can I download Podcast AI Edit now?
No public podcast release is available. Current public downloads provide the music-driven multicam workflow.
Does the test engine upload audio?
The prototype decodes and analyzes microphone audio locally. This does not remove the existing account and licensing requirements of the plugin.
Can one mixed audio track identify every speaker?
Not with the current test engine. It requires separate microphone tracks for 2–3 speakers.
When will it be released?
No date has been announced. Premiere integration and real-recording validation must be completed first.
Need music-driven multicam today?
The released R Project workflow analyzes beats, BPM and music drops for DJ sets, concerts and live events. Try that workflow with your own footage before purchasing.
DJ editing workflow · Panduan Bahasa Indonesia · Release notes