Separate the canvas from the footage
Your source dimensions describe the recorded frame. Your output canvas describes the shape you want to export. A 1920 × 1080 recording has a 16:9 ratio; a 1080 × 1920 canvas has a 9:16 ratio. Reversing those numbers does not rotate or recompose the subject. You have to decide what should happen to the image inside the new shape. This guide uses example canvases, not a statement of any platform’s current upload limits. Check the destination’s requirements when you export.
Use crop when filling the frame matters most
To fill a canvas without stretching the image, scale both source dimensions by the larger of the width and height scale factors. For a 1920 × 1080 source and a 1080 × 1920 output, the height determines the scale: 1920 / 1080 ≈ 1.778. The scaled image is about 3413.3 × 1920 pixels. A centered 1080-pixel-wide crop removes about 1166.7 pixels from each side of that scaled image. Only 31.6% of the original image area remains. A person centered in the recording might survive; a product sitting near an edge might not.
scale = max(output width / source width, output height / source height)
scaled width = source width × scale
scaled height = source height × scale
trim per side = (scaled width − output width) / 2Use fit when every edge needs to stay visible
Fit uses the smaller scale factor so the entire source remains inside the canvas. In the same example, 1080 / 1920 = 0.5625, so the image becomes 1080 × 607.5 pixels. That leaves 1312.5 pixels vertically, or 656.25 above and below when centered. Those spaces can remain bars or become a deliberate layout in your editor. The calculator reports the geometry and illustrates black padding; it does not generate a blurred background or export a finished video. Editors round fractional pixel positions when they render.
Compare the square version too
A 1080 × 1080 square crop of the same source uses a scale of 1. The crop trims 420 pixels from each horizontal edge and keeps 56.25% of the source area. A square fit instead scales to 1080 × 607.5 and adds about 236.25 pixels above and below. The landscape 1920 × 1080 canvas needs neither crop nor padding. These calculations show why a single automatic center crop is not a complete repurposing strategy: the amount of surviving context changes substantially between shapes.
1920 × 1080 source → centered crop
Portrait 1080 × 1920: 31.64% of source area remains
Square 1080 × 1080: 56.25% of source area remains
Wide 1920 × 1080: 100% of source area remainsInspect the opening, the widest gesture and the final frame
Before exporting, scrub the footage rather than judging only its thumbnail. Check a moment where the subject moves sideways, a frame containing a product label and the point where any text changes. Reposition the crop when necessary. Keep captions and essential labels inside the visible area, then inspect the destination’s own preview for interface overlays. If the source is a screen recording with small text, fitting may preserve the content while making it too small to read. Re-recording a closer view or editing a separate composition can be a better decision than forcing the original into every shape.
Export and review a short sample first
The calculator can flag enlargement, but it cannot measure whether the result looks sharp. A 1.778× enlargement adds output pixels without adding captured detail. Export a short representative segment, play it at its intended display size and check the subject, text and motion. Once the framing works, export the whole piece and review the final file. Then pair it with an opening line that matches the actual video and a tagged destination link if you are asking people to visit your website.