← ArchiveTalk2017
Autodesk Maya - Features - Delta Mush deformer
Official Autodesk feature demo of the Delta Mush deformer, which acts as a low-pass filter to remove skinning artifacts and restore rest-shape volume on deforming characters.
Abstract
This Autodesk feature demo introduces the Delta Mush deformer in Maya 2016, which reduces the need to paint skin weights by calculating offsets (deltas) between the reference mesh and a smoothed version and reusing them to remove deformation artifacts. The presenter deliberately smooth binds a character with bad skin weights set to a max influence of 1, scrubs the animation to expose ugly deformation across the body, then layers on the Delta Mush deformer and raises its envelope from 0 to 1 to show the artifacts cleaning up. The talk then adjusts the iterations and displacement attributes to control how strongly the smoothed deltas are applied, demonstrating how an otherwise unusable skin bind can be smoothed into a clean deforming character.
How to read this
- Category
- Production talk: a feature demo of the Delta Mush deformer
- Contributions
- Shows the Delta Mush deformer cleaning up artifacts from a deliberately bad smooth bind (max influence of 1) by raising its envelope from 0 to 1
- Demonstrates using stored offsets (deltas) between a reference mesh and a smoothed version to reduce skin-weight painting
- Walks through the iterations and displacement attributes to control how strongly the smoothed deltas are applied
- Context
- A vendor demo of the in-Maya implementation of Mancewicz et al.'s Delta Mush (Smoothing Deformations While Preserving Detail), presented as a practical low-pass filter for skinning artifacts.Builds on: Delta Mush: Smoothing Deformations While Preserving Detail
- Correctness
- Vendor feature demo, not peer-reviewed; the cleanup is shown on a single intentionally bad bind, so it illustrates typical behavior rather than validated limits, and very poor input or extreme poses may still need weight work.
- Clarity
- Very accessible; a single watch conveys the idea, no second pass needed beyond replicating the attribute settings.
- How to read it
- Watch for what the iterations and displacement attributes do and when Delta Mush substitutes for weight painting versus only masking issues; consult the source paper for the actual algorithm.
Built upon by
Nothing yet.
Related work
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →