SmockSim

Intention

I am interested in textiles and building computational design tools. Smocking is an interesting way to manipulate surface shapes, I wanted to see this can be mimicked computationally as well as explore hand smocking.

Process

A large number of smocking examples have the following flavor -- pleat the fabric using a running stitch, then gather pleats by sewing them together to hold the connection to create interesting surface textures.

After going through various hand smocking tutorials, I decided to begin with a very simple honeycomb pattern for hand smocking. During hand smocking, I realized that a sinusoidal grid mesh can be used to mimic a pleated surface instead of trying to simulate pleating. Woven fabric can be mimicked with a grid quad mesh with the constraint that all the faces try to remain planar quads. By adding a spring or closeness constraints between pleat points that are gathered together, the surface automatically relaxes to a shape that appears like a smocked fabric.

Product

I created a simulator to mimic a simple smocking technique. I also created a hand smocked honey comb pattern to build intuition to code up the simulator as well as verify the result.

The hand smocked prototype was created by sewing on a linen fabric. 

After drawing helper grid lines, a simple running stich was used to create pleats

Gathering consecutive pleats, alternating along rows gives rise to a honey comb pattern:

For the simulated version, I used a sinusoidal function to distort a grid as the base pleated mesh. Pairs of points can be selected to indicate that they will be sewn close together:

After selecting points to gather, I add closeness constraints on the gather points and planarity constraints on the quads to mimic smocking. The simulator attempts to iteratively relax the mesh by projecting the points to satisfy the constraints. After a few iterations, the mesh relaxes to the appropriate shape:

Reflection

A lot of fun and useful computational tools can be built around crafts where virtually experimenting can be useful. Hand smocking allows a large space of interesting designs where planning virtually is useful. Of course, capturing all the intricacies of the actual craft can be quite tricky. Particularly, I would like to extend my simulator to handle more styles like Canadian smocking patterns, deal with twisting pieces of fabric before sewing them and also have a good way to show embroidery.

Sample Book