Any type of image input may be used for loose and sketchy rendering, but some types of images work better than others. The main thing to remember is contrast. The edge detection process will find edges between any two areas that contrast with each other. So, try to make sure there is enough contrast to detect edges where you want them and little contrast in the areas that you don't want edges.

The following guidelines are suggested when creating or selecting input images.

Creating Depth Maps

When using depth map input, sometimes it may be necessary to "trick" the depth map renderer in order to create the best input.

Here's an example:

The desired image is one sphere slightly in front of another sphere. Here is the depth map, edge image, and loose and sketchy image created:

-> ->

The sphere towards the back has very little contrast with the background, which makes it hard for the loose and sketchy renderer to detect the edges around it.

To fix the problem, a plane is placed behind the spheres. It is made big enough so that its edges aren't seen in the image, but because it is there, the depth map renderer makes the second sphere a little darker.

-> ->

Creating a Depth Map in Alias|Wavefront Studio

In the "Render Globals", under "Image File Output", click "Depth" to render out a depth image for the desired camera view. Then render the image. Alias has its own format for depth information. To change the rendered depth file into an image useful for loose and sketchy rendering, use the following command:

Then convert the wrl image to an rgb or ppm file. For more on the depthmap utility, type "depthmap -h".

Creating a Depth Map in Maya

In the "Render Globals", under "Renderable Objects + Cameras", click "Depth" and "Renderable" to render out a depth image for the desired camera view. Also make sure that the "Image Format" is "Maya IFF". Then render the image using a batch render. Maya saves depth information in its own channel of the iff file, which then has to be extracted. To get the depth image useful out of the iff file into a format useful for loose and sketchy rendering, use the following command: fcheck depth_image

Type "z" to see the depth image and then "w" to save it as a wrl image. It is saved as "fcheck.out.1". Then convert the wrl image to an rgb or ppm file.

Automating the Procedure?

There are probably better ways to get the depth image from Maya and Alias, but I haven't done quite that much research on it at this time. If you know a better way to do this or have written scripts to automate the procedures, email them to me (abna@texas.net) and I will post it here.

Rendering Images for Use as Input

When using computer generated images other than depth maps as input into the loose and sketchy renderer, use strongly contrasting colors for the shaders to help the edge detection process find all of the right edges. If you don't want edges to be found around shadowed areas, use strong, flat lighting in the scene.

Examples of the results produced using different types of shading and lighting are shown below:

brightly lit with high contrast shaders
-> ->

brightly lit with low contrast shaders
-> ->

dimly lit with high contrast shaders
-> ->

dimly lit with low contrast shaders
-> ->

Using Other Types of Input

When using photographs, video, drawings, etc., first of all, remember the all-important contrast. Another thing to consider is to limit your subject matter. The more objects or patterns there are in the chosen subject matter, the harder it will be for the renderer to detect where one object ends and another begins.

Here are two examples:

A busy image:
-> ->

A simpler image:
-> ->

[Loose & Sketchy Rendering] [Disscussion] [User's Manual] [Examples & Tutorial] [Downloads]

[Depth Input] [Rendered Input] [Drawn Input] [Photo Input] [Tutorial] [Suggestions]