Skip to content

Squash and stretch animation with a free deformation lab

Learn squash and stretch using an original grounded ball study, constant 2D area calculations, a frame slider and downloadable SVG and CSV references.

FREE ANIMATION FRAME LAB

Inspect squash and stretch frame by frame

Original authored example · 24 saved frames at 24 fps · One second · No AI generation or footage upload

Original squash study at frame 12

Time 0.500000 seconds · Horizontal scale 1.6667 · Vertical scale 0.6000 · Area ratio 1

SVG contains the displayed drawing and any enabled onion overlays. CSV always describes the authored effect plan and ignores preview toggles. Coordinates use a 640 by 360 canvas with Y down. This is a local study, not a native animation project or a model benchmark.

Original round stretch and contact squash drawings with a fixed ground and constant 2D area

For the next production step, use the anticipation timing lab.

A soft ball landing on a floor should feel different from a steel bearing. Squash and stretch makes that difference visible through the drawing or rig. The useful question is not how far an object can deform, but what the deformation tells the viewer about its weight, material and action.

What is squash and stretch in animation

Squash compresses a shape along one direction while expanding it elsewhere. Stretch extends it while narrowing the other dimensions. Used with timing and clear poses, these changes can communicate an impact, a leap or an expressive gesture. Read Adobe's key-pose introduction for pose planning and the SideFX squash and stretch tool documentation for a rig-oriented implementation.

In this guide, the example is a flat ellipse, so the conserved quantity is drawn area. A three-dimensional object needs a volume-aware rig or deformation model instead. Keeping a 2D area constant does not establish that a 3D character preserves volume, muscle structure or material behavior.

Compare a rigid drawing with a deformed one

Watch this clip and view its details

Original silent authored comparison with one second of action and one second holding the final drawing

The left panel keeps a radius-40 circle. The right panel follows the same planned bottom-edge path but changes shape near contact. Source frame 12 has vertical scale 0.6 and horizontal scale about 1.666667. Frames 11 and 13 have vertical scale 1.3 and reciprocal horizontal scale. Other frames use unit scales.

The ground is Y = 300 in a 640 by 360 source canvas. We position the ellipse center using its current vertical radius, so changing the deformation does not push its bottom through the floor. This is an authored graphic trajectory rather than a physics simulation: it does not calculate mass, restitution or energy loss.

Preserve area without confusing it with volume

An ellipse has area π × horizontal radius × vertical radius. The lab uses horizontal scale = 1 / vertical scale, making their product one. A vertical scale of 0.6 therefore requires horizontal scale 1.666667 for the same 2D area.

StateVertical scaleHorizontal scaleArea ratio
Round111
Default contact squash0.61.6666671
Default adjacent stretch1.30.7692311

For a 3D ellipsoid, all three scale factors affect volume. If two lateral axes use the same scale, that scale is 1 / √vertical scale for constant volume. Do not apply the two-dimensional reciprocal rule to both lateral axes and assume the result conserves volume.

Use the free squash and stretch lab

  1. Inspect frame 12 to see contact against the fixed floor.
  2. Adjust contact deformation from zero to 0.6. Zero restores a round drawing.
  3. Compare frames 11, 12 and 13 to see stretch, squash and recovery.
  4. Turn deformation off to compare the rigid reference at the same frame.
  5. Save the displayed SVG and the complete timing plan.

The lab has 24 saved source frames at 24 fps. The final source frame starts at 0.958333 seconds and completes the one-second display interval. The video adds a one-second hold for inspection. Its background, position path and frame rate remain the same on both sides.

Download the 24-frame timing CSV, original renderer, drawing and timing source, reproduction notes and asset manifest. The source needs Bun, Sharp and FFmpeg. These are authored drawings and plans, not generated model results or native editor projects.

Apply the principle to a character or product brief

Decide which parts may bend before generating a shot. A fictional jelly character can compress heavily; a glass bottle should not squash just because the animation is energetic. Preserve labels and product geometry unless the concept intentionally transforms them. For a stylized mascot, approve its neutral silhouette and a contact pose as references.

The original soft teal mascot lands on the floor, briefly compresses at contact and recovers to its approved shape. Keep the eyes and costume recognizable. Preserve the floor and the mascot count. Use a readable short anticipation and recovery rather than repeated wobbling.

Adapt this untested direction to an operation available in Clout Image and Video or a sequence planned in AI Cinema. Exact deformation curves may require an animation rig or frame editing. A model response is something to inspect, not evidence that the requested geometry was followed.

Fix common squash and stretch problems

If the object seems to inflate, compare its outline at contact with the neutral pose. If it floats, inspect the bottom edge rather than the center point. If it feels rubbery when it should feel heavy, reduce the amplitude or shorten the recovery. If the entire body changes identity, simplify the shot and return to the approved reference.

Use onion skin animation to inspect neighboring positions. Use a smear drawing for a brief fast transition rather than extending every part of the action. Keep the material choice, trajectory and deformation rule in your shot notes.

Sources and testing

Original deterministic SVG drawings, frame arithmetic and native downloads tested by Clout. Adobe animation key-pose guidance and Blender onion-skin documentation checked October 24, 2026. AI production directions are original untested briefs, not output-quality benchmarks.

Questions, answered

Frequently asked questions

What does squash and stretch communicate

It can suggest weight, flexibility, impact and expression by changing shape in coordination with timing and poses.

Does the lab preserve three-dimensional volume

No. It preserves the drawn area of a two-dimensional ellipse. A 3D rig needs a volume-aware rule across all three axes.

Why does the ball not sink into the floor when squashed

The center height is adjusted using the current vertical radius, preserving the planned bottom-edge position.

Is the bounce a physics simulation

No. It is an authored trajectory and deformation study with no mass or collision simulation.

What does setting deformation to zero do

It restores horizontal and vertical scales to one for every frame, producing the round reference drawing.