Warp Morphing - Frames, Points and Links

Warp Morphing and Frames:

When we speak of frames with regard to warp morphing, we are referring to the concept that many morphs are intended to be animated. Because of this, the morphs would be performed over a series of intermediate frames where the effect is slightly more apparent in each of the frames. For instance, when someone says they are creating a 15-frame warp morph, the idea they are conveying is that in frame 15, the effect of the morph is fully apparent. In frame 8, the middle frame, the effect would be 50% of the final result, and so on for each intermediate frame. When a sequence of frames are rendered, the "flow" of the surface of the image under control of the points and links is apparent as the frames are viewed in series.

When the object of the morph is a single final image and not a sequence of images, the convention is to assume that the desired image is the final image of a sequence (of any length). As an example, if WinImages Morph has been set for a 17-frame sequence, rendering frame 17 will provide a result with the effect 100% apparent. WinImages Morph allows you to choose a frame from anywhere within the currently set sequence length when you render a single frame instead of a sequence of frames. Simply remember that when you are creating a single warped image you should set the frame to be rendered to be the same as the length of the sequence. This is done by selecting the Do 1 Frame option in the Generate menu.

Points:

Control points are used to define the flow of the image surface over the frames of a morph; when the object is one single warped image instead of a sequence, you should still think of the control points this way - it's just that you're only interested in the last frame of the sequence. The start point defines the portion of the image that will be moved, and the end point defines the location for that portion of the image for the final frame of the morph.

Control points can be thought of in two ways. First, and most conveniently, as a series of pairs of points (or lines), just as you see them in the main dual-window interface. Secondly, you can consider the two points as the endpoints of a path, over which the image must "flow" to accomplish the changes you are specifying with the points.

Placement of points should be done such that they are deposited at all key corresponding locations on the image surface. If we take the case of a face as an example, the corners of the eyes and mouth are obvious candidates. In addition, placement of points regularly along the edge of the profile will help control the image more exactly.

Points are never discarded by WinImages Morph's morphing procedures, even if they are defined in a conflicting manner. For this reason, you need to be careful that you do not place your points in such a way as to cause a conflict - one example of this is when the paths of two points cross each other during the course of a morph. This, and similar situations, will create folding and/or tearing in the image.

Points cannot completely restrain an image, because they only exactly control the portion of the image directly underneath where they are placed. Areas of the image which are not directly underneath a point (or an edge) are adjusted according to the motions of the points which are nearest. Even points which are quite far away may add motion to an otherwise uncontrolled region. When a portion of an image needs more exact control, you should use points in conjunction with links, described next.

Links:

Links are similar to points in that they specify movement for the surface of the image. They differ in that they control a much larger region, the entire area exactly under the line.

Links also differ in that WinImages Morph's morphing procedures may discard an edge under certain circumstances, such as when two links are defined that cross each other. When creating links, you should take care to ensure that you do not create a situation like this, as the results are undefined and will result in the image tearing or folding.

Placement of links should generally be done along lines in the image such as the links of eyes or a profile. They will restrict the motion of the image surface such that it will not transfer image data across the line, retaining colors precisely within the regions which are bounded by links. Most high-quality morphs will use almost as many links as they do points/lines.

Lines and Curves:

Lines and Curves are used to specify continuous regions of control and movement. The lines can be thought of as a nearly infinite amount of points along the specified line. This means that sections of a line can have different velocities and transparencies. Lines and curves are a quick and easy method of specifying control and movement in a morph, and in some cases, are easier to use than points and links. However, this does not mean that lines will produce superior quality over points and links. As with links, you will want to avoid crossing lines. Lines that cross will cause image folds or tearing.

Morphing Time:

The time it takes Morph to generate morph output frames is based almost entirely upon the size of the output images in the sequence. Anti-aliasing also adds a certain amount of time to the generation of each frame, again directly proportional to the size of the output image. A small amount of additional time is needed to compute frames that contain custom point-velocity computations. The size of the input images has no effect (and they may even be different resolutions, also with no effect on time); WinImages Morph always scales images as needed during operations.

WinImages F/x Manual Version 7, Revision 5, Level B