CAP 203 - Computer Animation III

Chapter 9: Rendering


This lesson covers material from chapter 9 of the text. Objectives important to this lesson:

  1. Rendering basics
  2. Render engines
  3. Other exercises

This chapter begins with a review of some basic knowledge about rendering. On pages 176 and 177, the author presents six things about rendering that you really want to know:

  • Most of the settings you want to use are in the Render Setup dialog box. Open it from the menu bar by selecting Rendering, Render Setup. Open it from the button bar by clicking the Render Setup button, which is just to the right of the Material Editor button. (She doesn't mention the button, but you should know it.)
  • The Time Output group (indicated in red on page 176), controls what frames will be rendered: one, all, a range, or a selection set.
  • The Output Size group controls the resolution of the rendered output. Check this before rendering to avoid creating something so big it crashes your workstation.
  • The Render Output group contains the Files button, which lets you select a name, file type, and destination for your rendered output. The author warns that you are not required to use this feature, but you must use it when rendering a movie or other type file, if you expect to have that file after the render process is over. No destination means no saved file.
  • The Show Safe Frames option places frame lines on a viewport that show what would be seen in a render at each of several resolutions. Use this to preview an animation before rendering it. Change the viewport (or move the camera) if you are not getting what you need.
  • See the tip in the margin of page 177. If you are rendering a file to print at a certain size and resolution, this is a good technique: multiply the desired dots per inch by the number of inches wide and inches high to get pixel resolutions for each of those dimensions. Use those pixel settings for the render output.

The second lesson (pages 178 and 179) discusses the Mental Ray renderer. You should read it to note how to switch from one render engine to another, and what the three engines are. The discussion of Mental Ray on pages 188 and 189 is more enlightening. Read it as well to understand what it does and why you will want to use it for some scenes.

Depth of Field Exercise

In this lesson we will consider the value of rendering part of the scene out of focus. The exercise begins on page 180.

  1. The illustrations in the text do not make the author's point as well as her discussion does. In real life, when you look closely at something close to you, you may see distant things that are behind it as well, but the distant objects will be out of focus and blurry. You are using a depth of field with your eyes in this example, focusing on what is nearby, and leaving what is distant out of focus.
    Open the start file for this lesson. Note that the objects in the scene are all in focus, as they are in all 3DS Max scenes. Render the scene as instructed and you will see the same thing in the rendered output.
  2. Examine the camera in the scene, checking the location of its target as instructed. You will need to zoom way in to see the satellite. Its size is not unreasonable compared to the size of the planet.
    Note that the target is a point in the scene, not necessarily an object in the scene. Enable the Multi-Pass effect for the camera, and select Depth of Field. (Or, if it is there, select Leeloo Dallas.) This will cause objects in the scene that are not at the target distance to be rendered less clearly.
  3. Select the camera, select the Camera02 viewport, and use the Preview function to see the effect of the default settings.
  4. Modify the Total Passes value and the Sample Radius value to make the effect more to your liking. Note the values that you choose. To get an effect I could see, I had to use a value greater than 1 for the Sample Radius. Do not choose a huge number for Total Passes.
  5. Render and learn what Total Passes really means. Render the scene to a file with the effect turned on and again with it turned off. Show me the files.

Rendering Large Images Exercise

This exercise is on pages 186 and 187. We have not done an exercise like this, so let's look it over. The author's story is interesting. Sometimes you want a program to do something it can't do. The best thing is to find a way to work around the problem. In this case, she wanted to render a file that was way too big. She found a way to render a piece of a view to a file, and to automate the process with a script (program) so she got all of the pieces of the view rendered as separate files.

  1. The author has not provided a scene for this exercise. You can use practically any scene you want. Assume you want to make the final render as she describes in the text: 15000 by 15000.
  2. She has rigged the numbers so that the final image is square and so are the pieces she is cutting it into. The pieces don't have to be square, but they do have to be the same, or the script file will not work. She determined (experimentally?) that she has enough RAM to render images that are 3000 by 3000, which will cut her final render output into five columns and five rows. She wants to make 25 frames that will each be 3000 by 3000 pixels. She will do this with a trick.
    Set the resolution for the scene to 3000 by 3000 on the Render Setup dialog.
  3. So the scene itself is now 3000 by 3000.
    We are going to carve 600 by 600 chunks out of the 3000 by 3000 scene, and force each chunk to render as a 3000 by 3000 frame.
    You need to know that the pixels in the scene are numbered using offset notation, starting with 0,0 in the upper left corner of the scene. The diagonally opposite corner of the first piece will be 599,599.
    The author explains the command in this step as rendering a section designated by coordinates to the full size of the render setup.
    render rendertype:#blowup region:#(0,0,599,599)
  4. There is a typo in the script that appears in the text. The variable in the first line is increment, not ncrement. This must be so, because the variable is used several times in the loop, and each of those times it is spelled increment.
    You can view the script provided on the DVD by clicking MAXscript, Open Script, then navigating to the script and clicking Open.
  5. In theory this will create a series of single frame render files that could be combined into the large image desired. In reality, trying it this way locked up my computer. It would have been nice to know the specs on her machine to guess what ours might be capable of doing. To do this realistically, we need to know how large an image our machine can render without crashing. Perhaps 3000 by 3000 is too ambitious for many computers.

    Update: 3000 by 3000 crashed in our classroom. It was possible to attain the desired final "resolution" by cutting the scene into many more smaller pieces.
    Desired output: 15000 by 15000.
    Set scene resolution to: 1500 by 1500.
    Set increment value to: 150.
    Change loops in script: run inner and outer loops 10 times, to create 100 reads of the scene, each 150 by 150, blown up to 1500 by 1500.