Orcast / Guide
THE ORCAST WORKFLOW
How to edit a multicam video podcast
From your shoot folder to a reviewed first cut. Follow each step to sync cameras by sound, set up automatic camera switching, and export your recorded podcast.
These are stylized workflow illustrations, not recordings of the current interface. Controls may look different in the app; each example can be played separately.
Import your podcast recordings
Drop your shoot folder onto New project, or choose the folder. In What we found, check the video and audio files. Remove anything that is not part of the shoot, use Add a file if something is missing, then press Next.
The example contains Guest.mp4, Host.mp4, Wide.mp4 and their audio recordings.
Set up the conversation
In Scenario, choose Two or more for a conversation. The microphone question appears below. Choose Yes, one each when each speaker has a separate recording.
Using one shared microphone? See the recording setups below for the speech-recognition path.
Match each close-up to its microphone
For the guest camera, choose Close-up, open Choose the sound, and select Guest.wav. Press Next and repeat for the host camera with Host.wav.
Match the person in the picture to their own audio track. File names in this example are illustrative.
Identify the wide shot
For a camera showing both speakers, choose Wide and select the corresponding sound recording. In this example, the room camera uses Stereo Mix.wav. Confirm the camera and continue.
A wide shot shows the conversation together; it is different from a close-up of one speaker.
Sync cameras by sound
Continue to Cut. Orcast aligns each microphone with its camera, then lines the cameras up using shared audio. Wait for synchronization to finish and review the aligned tracks.
Choose the pace of the edit and adjust the Cut settings if needed. Sync depends on the recorded sound; the progress shown here is illustrative.
Review the automatic camera cuts
Open Edit and play the timeline. Check that the picture follows the right speaker and that the sound stays in sync. Review interruptions, pauses and wide shots, then refine the cut before exporting.
The example switches between Camera 1 (guest), Camera 2 (host) and Camera 3 (wide).
Export MP4 or continue in your editor
In Export, choose Video (MP4) for a rendered file. Check resolution, frame rate and sound before exporting. Or choose Premiere Pro or DaVinci Resolve for a layered FCP7 XML timeline.
Keep the original recordings available when opening an XML timeline. FCP7 XML is not native Final Cut Pro FCPXML.
Match your recording setup
Automatic camera switching works best when Orcast can tell who is talking. That comes from the microphones, so the setup you recorded decides both the way it cuts and the settings. That is exactly what Orcast asks on the «Scenario» step — below is what to answer for your setup.
The classic podcast: each person on their own camera and their own microphone. Orcast hears who is louder and cuts to that person's close-up — automatic camera switching, in the desktop app.
Answer Two or more, then Yes, one each — Orcast takes the By voice path. On that same «Scenario» step, give each camera its clean mic so the switch follows the right person.
Two close-ups plus a wide that sees the whole room. The wide isn't tied to one speaker, so it isn't part of the by-voice switching — Orcast can insert it during long monologues and when people talk over each other, according to your Cut settings.
Answer Two or more and Yes, one each — that gives the By voice path. On «Scenario» set the room camera's framing to Wide, and on the «Cut» step set When one person talks and talks and When they talk over each other to go to it.
A talking head or a piece to camera. There's no one to switch to, so the cut changes on rhythm — landing on the pauses between phrases so nobody is cut mid-word.
Answer One, then where the cut should land: In the pauses, On sentence ends (that one needs the speech recognised) or Strictly on a timer. That is the By phrases path; on the «Cut» step you set how long a shot holds before it changes.
Everyone recorded on a single room mic or a stereo mix. A shared track sounds the same whoever is speaking, so it cannot tell people apart. Orcast therefore takes the By person path: who is talking comes from the recognised speech. It asks once which track that speech should be read from.
Answer Two or more, then One shared by all. Before it builds, press Recognise the speech — «By person» cannot be built without a transcript. Speech recognition runs over the internet only: the audio of the recording you picked goes to a cloud service — never the video. To switch cameras with no recognition at all, record a separate mic per speaker next time.
Want a tighter shot from a camera without a second lens? Add an extra close-up (crop) and drag the frame over the picture — Orcast crops that camera to it, for real, in the edit and the export.
On the Scenario step, press ＋ Extra close-up (crop) on the camera and position the frame.
Before recording your next episode
- Keep useful reference sound. Cameras and microphones need enough shared audio for sound-based synchronization.
- Record each speaker separately when possible. Separate microphone tracks support switching by voice.
- Keep cameras rolling. Gaps limit which shots are available for the edit.
- Keep your originals. Video stays on your computer. Building the cut requires an internet connection and sends technical audio data and editing settings to the online engine. Optional speech recognition sends selected audio to a cloud service.
Check your export before publishing
Watch a short section containing a change of speaker, an interruption and a pause. Check sync and framing in the MP4, or open the XML in your finishing editor and confirm the media is linked correctly. Recognized subtitles are saved as a separate SRT file; Orcast does not burn them into the video.
Editing in Premiere Pro · Editing in DaVinci Resolve · Podcast editing essentials
Try the workflow on your own recording
Start with a short section, check the result, and review the plans and limits for longer episodes.