[View the Image] [View the readme.txt file]

EMAIL: godfrey@shipbrook.net
NAME: Jeff Lee
TOPIC: School
COPYRIGHT: I SUBMIT TO THE STANDARD RAYTRACING COMPETITION COPYRIGHT.
TITLE: A Scholler's Tools
COUNTRY: USA
WEBPAGE: http://www.shipbrook.net/jeff/raytrace
RENDERER USED: POV-Ray 3.0 for OS/2 (unofficial compile)
TOOLS USED: OS/2 System Editor, Corel Draw 3.0, PhotoFinish, LView Pro 1.B/16
RENDER TIME: 27 hours  9 minutes  48.0 seconds
HARDWARE USED: Cyrix 6x86-166+ (32 MB, OS/2 Warp v4.0), HP ScanJet 4c
IMAGE DESCRIPTION:

This image is a portrayal of some of the materials needed by a young scholar (or, indeed, a mature scholar) around the turn of the seventeenth century, as mentioned in a short work entitled Rules made by F.B. for Children to write by. These Rules were contained in a book of penmanship published in 1611 by Richard Field. (A complete copy of the Rules may be found at <http://www.shipbrook.net/jeff/writing.html>, and explains, among other things, why a piece of bread is included.)

Some might object because the quill pen does not include the barbs of the feather, which are invariably left on the quill in "period" movies and television shows. I am not cheating by drawing only the calamus -- indeed, I am attempting to make the image more historically accurate; in reality, the barbs were removed (primarily to make the pen easier to handle), as instructed in the Rules:

The feather shaue off,the quill do not pare,
The stronger your pen in hand you may beare.

In other words, shave off the barbs, but don't cut into the shaft, as doing so will weaken the pen.

A more detailed description of the objects in the image is provided in the "readme.txt" file in the source archive.

(A note about the title: yes, I do know how to spell "scholar"; I merely chose to use early-17th-century orthography rather than modern, in keeping with the atmosphere of the image.)

DESCRIPTION OF HOW THIS IMAGE WAS CREATED:

The scene was designed by hand, using pencil sketches as a guide to composition and placement. The POV source files were created with the basic OS/2 text editor (i.e., no modellers were used).

The calligraphy was done by hand, and scanned with an HP ScanJet 4c. Using PhotoFinish, I manually converted the multiple tones of the paper to a single shade, so that I could use that palette index as the transparent colour for the image map.

Several objects (such as the lathe portions of the candlestick, the inkwell, and the Tudor-style house dimly visible through the window) were drafted in CorelDraw 3.0, in lieu of graph paper (by clicking on each of the individual points within the curves, I was able to easily ascertain the coordinates of each point, which I then typed into the source file -- tedious, but it works).

The meshes used for the brass candlestick were created with the aid of a hand calculator. It would have been easier to create only one set of facets and then create the others by using a #while loop and rotate; unfortunately, I didn't think of that until after I was finished.

The majority of the quill pen comprises a mesh that is created using nested #while loops; because most of the calamus has a cardioid cross-section, as well as a slight curve along its length, I opted to allow POV-Ray to do the calculations rather than doing them myself.

Although a few of the textures come from the standard POV-Ray include files (particularly the metal textures), the majority of them are my own creation. Colour values were arrived at by trial and error, though I did estimate several of them by changing a spare OS/2 folder's background colour and copying out the values (these are the ones that have RGB values like "209/255" rather than "0.8196" in the source).

The most challenging aspect of the image was in accurately reproducing the patterns of wear on the spines of the books. I used three books as the models (well, four really, but two of them were a two-volume set that were more or less identical), and each was made of a different kind of leather, with different amounts of damage, and different kinds of "ribs".

The first book that I modelled (book1.inc and book2.inc) had a doeskin spine that was incredibly smooth and somewhat hard. Small chips of the finished leather had been worn away, especially on the ribs and edges. This was fairly easily accomplished with a gradient y map, with each stratum (corresponding to an edge, rib, or inter-rib space) using a bozo pattern with a different ratio of scuffed/unscuffed leather and a different scale.

The second book (book4.inc) had an incredible amount of damage on the spine. Very few finished bits of leather remained, and two strata of suede-ish leather were showing. This was fairly simple to accomplish with an overall bozo pattern. Fortunately, the title had been rubbed off the original, so I didn't worry about how I was going to get it to show only on the finished bits.

The remaining book (book3.inc) scared me, so I left it for last. On this one, the spine had torn off from both the top and bottom, and there were cracks on the leather covering the centre bits of each rib, with unfinished leather showing on the top and bottom torn edges. The binding leather underneath was wrinkled where it had conformed to the shape of the quires (gatherings of pages), and was scuffed at both top and bottom.

I eventually ended up using a four-layers-deep special texture for the spine, which comprised several layers of special textures. Overall, it was a gradient along the Y axis, similar to that in book1.inc, but each of the segments corresponding to a rib was a gradient x texture which blended from a crackle texture at the centre (primarily unscuffed, with the scuffed texture in the cracks) to plain unscuffed leather at the edges. While the ribs on the other books were made of tori added to the spines, this book used tori differenced from the spine, and stretched along the Y axis to produce a curve up to the flat rib surface.

In retrospect, I should have moved the cork closer to the centre of the image, as the edge distortion produced by the perspective camera has distended it to the point that it looks as though it is intersecting the table, even though close-ups with an orthographic camera say it isn't. Unfortunately, there wasn't time to re-render, so I'm stuck with it.

I rendered using +a0.05, which yielded a fairly reasonable quality (although it increased the render time markedly). I used the unofficial OS/2 compile of POV-Ray 3.0 found on the povray.org ftp site.

When the image was finished rendering, I converted it to a JPEG using LView Pro with the quality set at 85%, and entropy optimisation enabled.

The image files containing the portraits, calligraphy and book pages were excluded from the source zipfile for reasons of space; please send me email if you would like a copy.


[View the Image] [View the readme.txt file]