Midokoro for Premiere
The whole clip, in another format
Turn a whole clip, or a stretch of it, into 9:16, 4:5, 1:1 or any format you choose, following whoever is speaking.
On the home screen, choose The whole clip, in another format. It's useful, for example, to turn a horizontal interview into a vertical version without having to move the frame by hand.
Choose the material
Transcript (optional)
Choose the stretch
Choose format and framing
Format
You can choose several formats at once: the first one analyzes the clip and the others reuse that analysis, so the next ones take almost no time. The formats are 9:16, 4:5, 3:4, 1:1, 4:3, 16:9 and Custom (a size of your own). Under the buttons you see the exact size in pixels for each one and a preview of how the crop looks.
Layout: how each scene is put together
| Automatic per scene | Shows one person at a time (whoever is speaking). Adds a second person only when it is clearly resolved, and shots without a clear face are fitted whole. |
| One subject | A frame that follows the person who is speaking. When the speaker changes, the frame jumps straight to the other person, with no sliding. |
| Two people | Both people at once, one on top and one below. Two people need to be on screen; if only one is found, it falls back to "One subject" and tells you. |
| Mosaic · up to 4 people | Everybody at once, each in their own quarter of the frame. It never repeats the same person in two cells. |
| Whole frame | The entire picture across the width, nobody cropped, with a Blurred background or Black bars. |
Following
| Automatic | Follows whoever is speaking, with a hard cut from one person to the other. |
| Fixed | A frame that stays still. You position it wherever you want afterwards in Premiere. |
| Centered | Always the middle of the picture, with no analysis. It's instant. |
With Zoom you choose how close the frame gets to the person: Close-up, Medium or Wide. It applies to one-person shots.
Cuts at every shot change
Turn on Leave the cuts made at every shot change and each clip arrives already cut where the shot changes in the video. That way you can adjust any shot on its own. It needs the analysis that follows the speaker.
What makes it slower
Reading who is speaking (the mouths) is the slow step. You'll see it marked with ⏱ slower: Automatic, Two people, Mosaic, the detailed check, and leaving the cuts made. Without those (One subject with fixed or centered framing) it is much faster and a bit less precise. Only the first format pays for the analysis: the others reuse it.
Detailed check of who is speaking
By default, Midokoro decides who is speaking using the audio and the position of the faces. When two or more people share the same shot and the audio isn't enough, turn on Detailed check of who is speaking (follows mouth movement): it is slower, but it tells who is moving their mouth. If each person appears in their own shot, the fast mode is enough.
Sequences
If the material is a sequence, Midokoro follows the picture down to the video file on the track below. Stretches whose picture can't be traced to a file (a nested sequence, a speed change, or a missing file) are framed centered and the panel lets you know. To follow the speaker in those cases, choose the clip from Project clips.
When it finishes, the result appears in your project: see What ends up in your project. With Do another you start over.