1. This article explores the basic challenges, concepts, and methods for motion editing, specifically on motions created by motion capture.

2. Why motion editing
1). Clean-up
2). Re-use
3). Creating Infeasible Motions
4). Imperfections of reality
5). Change of intent
6). Secondary motion

2.1. Unique Challenges of Motion Capture Data Editing
1). The data is most certainly inconvenient for editing.
2). There is nothing but the data to describe the properties of the motion.

3.  Representation of Poses and Motions
3.1. Hierarchical representation: a character is made up of a set of rigid segments  that are required to remain connected. A skeleton can be parameterized by the position and absolute orientation of one of the pieces, and the relative orientations between connected pieces.

3.2. Representation of a 3D rotation
1). Euler Angles
2). Axis/Angle
3). Quaternions
4). Exponential Maps

3.3. Representation of motion: A motion is a function that converts from a time to a pose. Usually, motion is a pose sequence.

4. Cleaning Motion Data: the process of making the data from our motion capture process more accurately reflect the performance  trying to capture.

5. Types of Motion Editing Techniques
5.1. Given motion curves, the problem of motion editing is simply changing these curves to meet our new needs, while preserving what we don¡¯t want to change.

5.2. Strategies to bridge the gap between the high level descriptions we would like to use, and the low level details we must control:
1). describe desired features by example.
2). control and maintain certain mathematical properties in the motion with hopes that preserving these properties can preserve the higher-level motion properties.
3). identify and control details in the motion, for example footplants, and make sure that these are maintained.
4). try to make editing the basic details more convenient so that it is easier to make the changes to the parameters.

5.3. Approaches to address the pragmatic issues of handling the large amount of unstructured data.
1). convert the data to a more convenient representation.
2). create methods that describe changes in ways that are independent of the underlying representation of the curves.

6. Key Reduction:  reduce the number of specified frames to get editable motion
The basic idea is to find curves that have simple representations, but fit the original data. Mathematically, this problem is called curve fitting or regression.

7. Motion Signal Processing: apply the theory and methods of signal processing to motion editing
7.1 Simple Time Transformations
1).Time scaling
2). Time warping

7.2. Filtering

7.3. Additive Motion Editing: Specifically, blending

7.4. Motion Warping
1). find desired changes on individual frames.
2). determine the displacement values at each of the changed frames by subtracting the original from the new poses.
3). construct a motion displacement map (a.k.a. a motion warp) that interpolates these known displacements.
4). compute the final motion by adding the original motion and the computed displacement map.