Table of Contents

1. Getting the program
2. Typical system requirements
3. Installation, libraries and other required things
4. Running POV-Ray

Getting the program

POV-Ray releases are available from the official POV-Ray website, www.povray.org. There are ready-to-run packages, with a compiled POV-Ray program and all necessary support files, for MS-DOS, Win95/98/NT, Macintosh, Linux, and Sun (Unix). There is also a full source-code package, in C, for people who want to compile POV-Ray on their own.

The current version of POV-Ray is 3.1. I am, however, most familiar with POV-Ray 3.0, which is similar except for some advanced features. With a few very minor exceptions (which I'll get to when we come to them), all of the descriptions, examples, and scene files I describe here work with any POV-Ray version 3.0 or higher.

Typical system requirements

Because of the way ray-tracing works -- simulation of the paths light rays take to each pixel of the image -- it is computationally very expensive. For all but the simplest renditions, POV-Ray runs very slowly on computers less powerful than the first generation of Pentium-based PCs and Power Macintoshes. A Pentium II-based PC or G3 Power Mac is much better still.

A little history: As an example, back in 1998 I used a Silicon Graphics Octane with 256 MB of memory and a 175 MHz IP30 processor (R10000 CPU and R10010 floating-point unit) to produce the graphics on this web site. A new (~300 MHz) Pentium II runs POV-Ray similarly quickly. By comparison an older Silicon Graphics workstation (an Iris Indigo) runs POV-Ray about 10 times more slowly, and a Macintosh IIsi with a math coprocessor (roughly equivalent to a 386-based PC) about 100 times more slowly.

Recent editorial note: By contrast, I recently (early 2005) re-rendered a complex scene I submitted to the IRTC in 1998. Rendered on the SG Octane, it took most of a night; on my laptop in 2005, it took about 30 minutes.

Since most of the work POV-Ray does is mathematical, floating-point unit (math coprocessor) speed is much more important than graphics and display processing speed.

Installation

Assuming you have a computer that can run one of the precompiled packages, just download the package, uncompress it, and follow the installation directions. Compiling POV-Ray yourself is possible, but it can be difficult, and is outside the scope of t his document anyway.

There are a number of important parts of the package to pay attention to. At some point, you may need to specify a location for the library of "include" files. Include files contain, among other things, a large number of pre-defined terms for things like colors, textures, and shapes, which will come in extremely handy when you are actually writing POV-Ray files. Generally speaking, the POV-Ray include files will go in a directory / folder called "include", in the same directory as the actual POV-Ray executable file.

Running POV-Ray

POV-Ray, by itself, has no graphic interface. Its sole job is to take a POV-Ray input file, also called a scene file, and render it. You write the scene file, using the POV-Ray language, which we'll discuss in greater detail later. The result of all of this is a bitmapped image. As a result, there are three things you must do to use POV-Ray at all:

1. Get a good image viewing program. There are many free/shareware programs out there, e.g. xv for Unix, GraphicConverter for Macintosh, and so on. POV-Ray's default output format depends on the platform. On Macs, it makes a PICT file; on Windows, it makes a BMP file. In all cases the format is a 24-bit color format, which provides extremely good color rendition.
2. Write a scene file using the POV-Ray language (discussed in more detail in later parts of this site). You use a text editor for this -- many versions of POV-Ray come with an editor built-in.
3. Tell POV-Ray details about how the input file is to be processed. You must tell POV-Ray several things, such as the name of the input file, the size and quality of the final image, and so on. As a result, getting POV-Ray to do what you want can get a little tricky, especially with a command-line interface such as MS-DOS or many Unix versions.

The most important "quantities" to set are usually the following:

• Input file name
• Output file name
• Width and height of the final image
• Locations of any special files POV-Ray needs

On a GUI operating system (e.g., Win95 or Macintosh) you set these from a menu. The documentation supplied with the program will specify which menus / screens will give you access to these parameters.

On command-line interfaces (MS-DOS and many Unix executables), you set these values on the command line, along with the name of the program, with "switches" that consist of either a plus ("+") or minus ("-") sign followed by the value in question. For switches that merely set a value, like the width of the image, either sign may be used, but for switches that "toggle" a feature on or off, "+" turns the feature on while "-" turns it off.

• +Iinput file name The name of the POV input file, which usually has the extension .pov. E.g. +Ipicture.pov. Note that the file name gets attached directly to the +I" flag. If you don't specify an extension, POV-Ray will use .pov.
• +Ooutput file name The name of the output (rendered) file.
• Wnumber The width* of the output file, in pixels.
• +Hnumber The height of the output file, in pixels.
• +Ldirectory path The location of the library of include files**, where the directory path is with respect to the location of the POV-Ray executable. Include files contain useful pre-defined terms that can greatly simplify writing POV-Ray files. E.g., type "+Lsample/include" if you have some include files located in a dire ctory called "sample/include".
• +D Turns on a crude display of the image as it renders. This is handy if you want to monitor how the rendering is proceeding. -D will turn off image display.
• +A Turns on anti-aliasing, a means of smoothing out lines that look jagged in a bitmapped image. Anti-aliasing is off by default. Anti-aliasing slows down rendering, so you'll only want to use it for a final render.

 

Left, anti-aliasing off; right, anti-aliasing on

Suppose you want to render a POV-Ray file on a Unix version. You want to use dimensions 320 by 240 (which will fit well on a typical Web page), and an input file called "monitor.pov". The default output format in the Unix version is Targa (.tga). Thus, you'd type:

povray +W320 +H240 +Imonitor.pov +Omonitor.tga

All versions of POV-Ray come with sample input files. Try running POV-Ray on several of these!

[ Index | Part 1 | Part 2 | Part 3 | Part 4 | Part 5 | Part 6 | Part 7

*Normally, POV-Ray works with an aspect ratio (width:height ratio) of 4:3, so you'll usually want to use values like +W320 +H240 or +W1024 +H768. If the width and height aren't in a 4:3 ratio the image produced will look distorted. It's possible to change the aspect ratio, but this generally won't be necessary for what we plan to do.

**A normal installation will set up POV-Ray to recognize the default "include" subdirectory. Also, POV-Ray automatically checks the current directory (usually, the one with the executable in it) for include files. Files in other pla ces normally need to be specified with +L.

There is also a file called POVRAY.INI, which contains a place for specifying directory paths to directories containing include and other library files. If these paths are specified in POVRAY.INI, you don't need to specify them on the command line wit h the +L flag. If you do a lot of POV-Ray work, or put include files in unusual places, editing the POVRAY.INI file will save you a lot of typing!