Midokoro

Formats and speaker tracking

The same clip, arranged for each network. Midokoro does not just crop the center: it follows who is talking.

Create versions

Press Crear versiones (Create versions) on a clip, or on the whole project. Tick the formats you want: 16:9, 9:16, 4:5 and 1:1. The original 16:9 is cut without reprocessing the video, so it is instant. The others are rendered and take a few seconds each; a compatible graphics card speeds it up. You can close the window: it keeps working in the background. When it is done, Descargar todas (.zip) (Download all) gives you every version.

The switch Subtítulos grabados en el video (Subtitles burned into the video) puts the subtitles on the picture, rearranged for each format inside a safe zone.

How to arrange the picture

Seguir al que habla (Follow the speaker)	Detects the scenes and who is talking. One fixed shot per scene, with a straight cut when the speaker changes.
Dos personas, arriba y abajo (Two people, top and bottom)	When two people are on screen, each one is framed in their half. With only one, it follows that person.
Automático por escena (Automatic by scene)	Proposes the best layout for each scene: one subject or two people.
Completo con fondo borroso (Full, blurred background)	The whole video, with a blurred background filling the rest.
Completo con barras (Full, with bars)	The whole video, with bars at the top and bottom.
Recorte fijo a mano (Fixed crop by hand)	Fills the frame. Drag over a preview to choose which part stays.

How speaker tracking works

Midokoro detects the scenes of the clip (the shot changes) and, inside each one, the people and who is speaking, from the voice and the faces. Each scene gets a fixed shot, and the shot only changes with a straight cut when the scene or the speaker changes. The camera never pans or zooms.

Who is speaking is an estimate. A microphone that covers a mouth, or people talking over each other, can confuse it. When it gets one wrong, correct it by hand as explained below.

Correct the framing by hand

In the versions window press Ajustar a mano (Adjust by hand). The Reencuadre (Reframe) screen opens. If the clip has not been analyzed yet, press Analizar ahora (Analyze now): it takes about 10 to 20 seconds.

Pick the format at the top (9:16, 4:5 or 1:1) and a scene in the strip of scenes. Dividir acá (Split here) cuts the scene at the playhead and Unir con la siguiente (Merge with the next) joins it to the next one.
Choose the shot of the scene: Un sujeto (One subject), Dos personas (Two people) or Completo (Full, with Fondo borroso or Barras negras).
Under A quién seguir (Whom to follow) choose Automático (the one who is talking) or a person.
Drag the box over the picture, pull a corner or use the mouse wheel to zoom. The arrow keys move it (hold Shift to move more) and + / - zoom. Tab switches between the top and bottom box.
Keys: Space plays or pauses, [ and ] go to the previous or next scene, 1 2 3 pick 9:16, 4:5 or 1:1, A returns to automatic and Esc closes.
Volver a automático (Back to automatic) undoes your change in that scene, and Reanalizar (Re-analyze) detects everything again. Changes are saved as you make them.

Scenes you adjusted by hand are respected when you re-analyze and when you create versions.

Your first project Editor and export