Fix frame alignment without flattening the motion

Use a fixed canvas and a previous-frame overlay to inspect unwanted jumps while keeping deliberate movement.

Consistent frame dimensions and a stable reference point help an animation play without accidental jumps. In FrameMint every exported frame shares a canvas. Position adjustments move the existing pixels; they do not redraw a pose or identify a character automatically.

Decide what should stay still

For a running-in-place character, choose a useful reference such as the standing foot, pelvis or center of the body. For a jumping character, vertical movement may be intentional. Aligning every silhouette to the bottom would remove that movement, so compare the action rather than blindly matching a bounding box.

Compare neighboring frames

Pause the animation and select the second frame. Turn on “Previous-frame overlay”. The preceding frame appears faintly behind the selected one. Adjust Move X and Move Y by a small number of pixels, then play the full cycle again. Positive X moves right; positive Y moves down. The first frame has no preceding-frame overlay.

Keep room for the widest pose

The canvas remains the size of the grid cell, or the maximum source width and height for a PNG sequence. Moving a frame can push a hand or effect past an edge and clip it. Check the preview before export. If the original sheet cuts off a limb between cells, offsets cannot restore those missing pixels. Correct the source image or use a larger source cell.

Be careful with trimming and transparency

Transparent padding can preserve a shared coordinate system. Independently trimming each pose and then centering it can change the apparent position. FrameMint keeps the fixed canvas and does not automatically trim each pose. Multiple PNGs of different sizes start centered on a common canvas, so inspect their alignment manually.

Removing a plain background

If every cell has one flat background color, enable “Remove a solid color” and choose that color. Tolerance is the maximum difference per RGB channel, from 0 to 100. Matching pixels are removed wherever they occur, including inside a character. This is a color key, not semantic background segmentation. A painted checkerboard or complex background cannot be reliably removed with this switch.

Return to the GIF workflow to check the final loop, or keep alpha transparency with a PNG export.

By Draconova · Updated 14 September 2026. This guide describes the current tool; use the preview and check your downloaded output.