From: mconst@soda.CSUA.Berkeley.EDU (Michael Constant)
Newsgroups: alt.fan.warlord
Subject: Re: Poorly Hung celebrities
How to Make a 3D BUAF Properly
The reader may already be aware of the profusion of 3D BUAFs in sig files.
No one seems to take into account the fact that a 3D BUAF is much more
difficult to design than a flat BUAF, and yet more and more people are using
the 3D BUAF in sigs and plans. It is not the place of this humble writer
to decide whether the 3D BUAF is, in fact, superior to the traditional flat
BUAF. However, many 3D BUAFs designed by amateurs fail to obey certain
laws of physics and of aesthetics, and I have taken it upon myself to
illuminate some of the finer points of 3D BUAF design.
For an example, let us take Joe Flannigan's rendering of "Joe" (figure 1.1).
Joe Flannigan wrote:
>--
> ____ _____________ __________
> |\::::\ |\::::::::::::\ |\::::::::::\
> |##\::::\ |##\::::\~~\::::\ |##\:::\~~~~~ __________
> ____ \###\::::\ \###\::::\ \::::\ \###\::::::\ \ Colour /
>|\::::\__\###\::::\ \###\::::\__\::::\ \###\:::\_______ \ This /
>|##\::::::::::::::::\ \###\::::::::::::\ \###\::::::::::\ \ Pi /
> \##|~~~~~~~~~~~~~~~~| \##|~~~~~~~~~~~~| \##|~~~~~~~~~~| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.1: the amateur 3D BUAF
Perhaps the most riveting defect is the use of the # character for the
shadow sections. The reader will notice that the # is being used to fill
a left-slanted region; however, the # is normally a right-slanted or
unslanted character. The unsuitability of the # is therefore apparent.
A better choice would have been the left-slanted % (figure 1.2). Notice
how the shadow regions seem more contiguous.
> ____ _____________ __________
> |\::::\ |\::::::::::::\ |\::::::::::\
> |%%\::::\ |%%\::::\~~\::::\ |%%\:::\~~~~~ __________
> ____ \%%%\::::\ \%%%\::::\ \::::\ \%%%\::::::\ \ Colour /
>|\::::\__\%%%\::::\ \%%%\::::\__\::::\ \%%%\:::\_______ \ This /
>|%%\::::::::::::::::\ \%%%\::::::::::::\ \%%%\::::::::::\ \ Pi /
> \%%|~~~~~~~~~~~~~~~~| \%%|~~~~~~~~~~~~| \%%|~~~~~~~~~~| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.2: correcting the shadow
The reader's attention will then be drawn to the impossibility of the
3-dimensional shapes of the O and E. The O has no interior, and the
middle and top horizontal segments of the E have no thickness. The former
problem can be easily remedied (figure 1.3).
> ____ _____________ __________
> |\::::\ |\::::::::::::\ |\::::::::::\
> |%%\::::\ |%%\::::\~|\::::\ |%%\:::\~~~~~ __________
> ____ \%%%\::::\ \%%%\::::\ \::::\ \%%%\::::::\ \ Colour /
>|\::::\__\%%%\::::\ \%%%\::::\__\::::\ \%%%\:::\_______ \ This /
>|%%\::::::::::::::::\ \%%%\::::::::::::\ \%%%\::::::::::\ \ Pi /
> \%%|~~~~~~~~~~~~~~~~| \%%|~~~~~~~~~~~~| \%%|~~~~~~~~~~| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.3: fixing the interior of the O
Note how by the addition of a single |, the O has gained a substantially
more realistic look.
Correcting the E is more difficult, since some of the space which should
be taken by front faces is already full. It is tempting to simply add
vertical bars and front faces where necessary (figure 1.4).
> ____ _____________ __________
> |\::::\ |\::::::::::::\ |\::::::::::\
> |%%\::::\ |%%\::::\~|\::::\ |%%\:::\~~~~~| __________
> ____ \%%%\::::\ \%%%\::::\ \::::\ \%%%\::::::\ \ Colour /
>|\::::\__\%%%\::::\ \%%%\::::\__\::::\ \%%%\:::\~~|____ \ This /
>|%%\::::::::::::::::\ \%%%\::::::::::::\ \%%%\::::::::::\ \ Pi /
> \%%|~~~~~~~~~~~~~~~~| \%%|~~~~~~~~~~~~| \%%|~~~~~~~~~~| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.4: the wrong way to fix the E
However, it is apparent from the illustration that the loss of the _ borders
has dramatically affected the appearance of the bottom horizontal segment
of the E. A better solution is to remove the top shading of the E, thus
allowing for more use of the _ character (figure 1.5).
> ____ _____________ ___________
> |\::::\ |\::::::::::::\ |\ ______\
> |%%\::::\ |%%\::::\~|\::::\ |%%\ \____| __________
> ____ \%%%\::::\ \%%%\::::\ \::::\ \%%%\ __\ \ Colour /
>|\::::\__\%%%\::::\ \%%%\::::\__\::::\ \%%%\ \_|______ \ This /
>|%%\::::::::::::::::\ \%%%\::::::::::::\ \%%%\ \ \ Pi /
> \%%|~~~~~~~~~~~~~~~~| \%%|~~~~~~~~~~~~| \%%|~~~~~~~~~~| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.5: another try at fixing the E
However, this does not work either, since the | of the middle segment
pierces the ________ row marking the top of the lower segment. The cure
is to use different characters to indicate different slants, and thus
different faces (figure 1.6).
> ____ _____________ ___________
> |\::::\ |\::::::::::::\ |\ ______\
> |%%\::::\ |%%\::::\~|\::::\ |%%\ \____| __________
> ____ \%%%\::::\ \%%%\::::\ \::::\ \%%%\ __\ \ Colour /
>|\::::\__\%%%\::::\ \%%%\::::\__\::::\ \%%%\ \ | \ This /
>|%%\::::::::::::::::\ \%%%\::::::::::::\ \%%%\ ````````\ \ Pi /
> \%%|~~~~~~~~~~~~~~~~| \%%|~~~~~~~~~~~~| \%%|""""""""""| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.6: the right way to fix the E
Here, the ` and " characters have been used to indicate the upper and
front surfaces of the E, respectively. Of course, since the E has lost
the shading of its top surface, the other letters must be made to match
(figure 1.7). The choice of a shaded or unshaded top surface is a
personal matter and this author will not presume to suggest one over the
other. However, because of the format of the E here, a shaded top surface
is simply not feasible.
> ____ _____________ ___________
> |\ \ |\ \ |\ ______\
> |%%\ \ |%%\ \~|\ \ |%%\ \____| __________
> ____ \%%%\ \ \%%%\ \ \ \ \%%%\ __\ \ Colour /
>|\ \__\%%%\ \ \%%%\ \__\ \ \%%%\ \ | \ This /
>|%%\ \ \%%%\ \ \%%%\ ````````\ \ Pi /
> \%%|""""""""""""""""| \%%|""""""""""""| \%%|""""""""""| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.7: matching the other letters
Note that the " character has been introduced into the J and O in order
to match the E.
Unfortunately, the loss of shading has made apparent several other problems
which had been less visible before. The worst of these is the lack of a
border between the left-hand face of the large vertical segment of the J
and the upper face of the J. This can be fixed with the same technique as
was just used with the E (figure 1.8).
> ____ _____________ ___________
> |\ \ |\ \ |\ ______\
> |%%\ \ |%%\ \~|\ \ |%%\ \____| __________
> ____ \%%%\ \ \%%%\ \ \ \ \%%%\ __\ \ Colour /
>|\ \ \%%%\ \ \%%%\ \__\ \ \%%%\ \ | \ This /
>|%%\ ```````` \ \%%%\ \ \%%%\ ````````\ \ Pi /
> \%%|""""""""""""""""| \%%|""""""""""""| \%%|""""""""""| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.8: finished?
This certainly looks much better than the original, and some BUAF artists
might even stop at this point. However, the mark of the professional is
the ability to notice and correct even the smallest of mistakes. And here
we notice that while the J is better than before, there is still some lack
of depth in the left-hand face of the large vertical segment. An effective
cure for this can be had by redoing all the left-hand faces in what is
known as the "segmented slant" style (figure 1.9).
> ____ _____________ ___________
> |\ \ |\ \ |\ ______\
> |\|\ \ |\|\ \~|\ \ |\|\ \____| __________
> ____ \|\|\ \ \|\|\ \ \ \ \|\|\ __\ \ Colour /
>|\ \ \|\|\ \ \|\|\ \__\ \ \|\|\ \| \ This /
>|\|\ ```````` \ \|\|\ \ \|\|\ ````````\ \ Pi /
> \|\|""""""""""""""""| \|\|""""""""""""| \|\|""""""""""| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.9: the "segmented slant" style
This style is considered radical by many of the more established BUAF
artists, but I believe that it has its place. Certainly the J has
benefitted from it, and the other letters do not seem to have suffered.
In redoing the left-hand face of the E, it became apparent that the left-
hand face was, in fact, misaligned in the original (see figure 1.1).
While the defect was not noticeable with the old shaded faces, it would
have been painfully obvious with the segmented slant faces, so I have taken
the liberty of correcting it without providing a seperate illustration.
However, the realignment of the E brings up an interesting point. It is
generally accepted that when a _, | and \ meet, the "tight intersection"
is to be preferred over the "loose intersection" (figuer 1.10).
> ____\ ____\
> | |
>
> tight intersection loose intersection
Figure 1.10: tight vs. loose intersection
However, the tight intersection in the middle of the E looks cramped
because of the extra \ to the left. An inexperienced BUAF artist might
be inclined to use loose intersection, but he would be overlooking the
better possibility of making the segments of the E longer (figure 1.11).
> ____ _____________ ____________
> |\ \ |\ \ |\ _______\
> |\|\ \ |\|\ \~|\ \ |\|\ \_____| __________
> ____ \|\|\ \ \|\|\ \ \ \ \|\|\ ___\ \ Colour /
>|\ \ \|\|\ \ \|\|\ \__\ \ \|\|\ \ | \ This /
>|\|\ ```````` \ \|\|\ \ \|\|\ ````````\ \ Pi /
> \|\|""""""""""""""""| \|\|""""""""""""| \|\|""""""""""| \nk/
> \|________________| \|____________| \|__________| \/
>
Figure 1.11: loose intersection is avoided
The tight intersection is maintained here, but the jumble in the center
of the E has been dealt with.
Finally, a few cosmetic changes need to be made: the ~ in the O should be
replaced by a " to indicate which surface it represents, and the upper
borders of the J need to be lengthened. The final product is shown in
figure 1.12.
> _____ _____________ ____________
> |\ \ |\ \ |\ _______\
> |\|\ \ |\|\ \"|\ \ |\|\ \_____| __________
> _____ \|\|\ \ \|\|\ \ \ \ \|\|\ ___\ \ Colour /
>|\ \ \|\|\ \ \|\|\ \__\ \ \|\|\ \ | \ This /
>|\|\ ```````` \ \|\|\ \ \|\|\ ````````\ \Pert/
> \|\|""""""""""""""""| \|\|""""""""""""| \|\|""""""""""| \h!/
> \|________________| \|____________| \|__________| \/
>
Figure 1.12: the finished product
Here is the finished product. While not comparable to the 3D BUAFs created
>from scratch by professionals, it is nevertheless a marked improvement over
the naively drawn first attempt it sprung from. The process of refining
an improperly-rendered BUAF is important to understand, because it is not
always possible to redesign a faulty BUAF from scratch. Often one must, as
I have done here, start with something decidedly amateur and create from it
a proper 3D BUAF.
While few high-quality 3D BUAFs are created nowadays, with knowledge and
perseverance a BUAF of distinction can always be created. While I cannot
sanction the creating of any BUAF, I hope that should the reader choose to
create his own 3D BUAF, this document will have proven helpful.
===================================
From: slavins@entergrp.demon.co.uk (Simon Slavin)
mconst@soda.CSUA.Berkeley.EDU (Michael Constant) wrote:
> [much snippage]
> > _____ _____________ ____________
> > |\ \ |\ \ |\ _______\
> > |\|\ \ |\|\ \"|\ \ |\|\ \_____| __________
> > _____ \|\|\ \ \|\|\ \ \ \ \|\|\ ___\ \ Colour /
> >|\ \ \|\|\ \ \|\|\ \__\ \ \|\|\ \ | \ This /
> >|\|\ ```````` \ \|\|\ \ \|\|\ ````````\ \Pert/
> > \|\|""""""""""""""""| \|\|""""""""""""| \|\|""""""""""| \h!/
> > \|________________| \|____________| \|__________| \/
>
> Figure 1.12: the finished product
You didn't share the right-hand-side of the inside of the 'O', Mike.
Simon "Is there a hyphen in 'anal retentive' ?" Slavin
========================
From: mconst@soda.CSUA.Berkeley.EDU (Michael Constant)
Simon Slavin wrote:
>mconst@soda.CSUA.Berkeley.EDU (Michael Constant) wrote:
>>[much snippage]
>>> _____ _____________ ____________
>>> |\ \ |\ \ |\ _______\
>>> |\|\ \ |\|\ \"|\ \ |\|\ \_____| __________
>>> _____ \|\|\ \ \|\|\ \ \ \ \|\|\ ___\ \ Colour /
>>>|\ \ \|\|\ \ \|\|\ \__\ \ \|\|\ \ | \ This /
>>>|\|\ ```````` \ \|\|\ \ \|\|\ ````````\ \Pert/
>>> \|\|""""""""""""""""| \|\|""""""""""""| \|\|""""""""""| \h!/
>>> \|________________| \|____________| \|__________| \/
>>
>> Figure 1.12: the finished product
>>
>
>You didn't share the right-hand-side of the inside of the 'O', Mike.
*ahem*
As you will no doubt recall from my earlier posting, the reason for the use
of the segmented slant style was that the J had lost some of its sense of
depth without it. The application of segmented slant to the O was purely
for consistency's sake, and it was certainly not for the sake of helping the
O's interior. And as anyone with even moderate familiarity with standard 3D
BUAF design would be aware, the use of the segmented slant style involves
only the major surfaces -- the minor surfaces are left untextured. I had
adhered to this usage, because the use of segmented slant on minor surfaces
carries with a host of problems (figure 1.13).
> _____ _____________ ____________
> |\ \ |\ \ |\ _______\
> |\|\ \ |\|\ \"|\ \ |\|\ \_____| __________
> _____ \|\|\ \ \|\|\ \\|\ \ \|\|\ ___\ \ Colour /
>|\ \ \|\|\ \ \|\|\ \__\ \ \|\|\ \ | \ This /
>|\|\ ```````` \ \|\|\ \ \|\|\ ````````\ \Pert/
> \|\|""""""""""""""""| \|\|""""""""""""| \|\|""""""""""| \h!/
> \|________________| \|____________| \|__________| \/
Figure 1.13: a rank beginner attempts to help
As is clear from this illustration, the segmented slant, if it were carried
any further, would collide at the bottom of the inner right-hang face with
the __ below it. The way it is shown above looks silly, because (among
other things) the segmented slant appears to stop halfway down the inner
right-hand face of the O. Of course, one could continue the segmented slant
style all the way down the inner right-hand face at the expense of the
aforementioned border, but this would destroy the 3-dimensional quality of
the inside of the O. And it is apparent from figure 1.13 that even were a
way found around this problem, the interior of the O would still look far
worse with the segmented slant than without, since texturing on the interior
face would appears to meld with the edges of the other faces, destroying any
sense of depth. It should be obvious that the sense of depth is the primary
reason for using a 3D BUAF, as opposed to a flat BUAF, and so it would be a
cardinal mistake to do anything which so drastically disturbs that quality.
If you have any further "suggestions", I would appreciate it if you first
consulted one of the well-known BUAF authorities before wasting my time with
the matter. I am sure some of my distinguished colleagues in this group
would be glad to refer you to the standard textbooks if you are having
difficulty understanding any other part of the analysis I posted earlier.
==========================
From: silver@eden.com (silver)
Michael Constant (mconst@soda.CSUA.Berkeley.EDU) wrote:
: How to Make a 3D BUAF Properly
[ much snippage ]
: > _____ _____________ ____________
: > |\ \ |\ \ |\ _______\
: > |\|\ \ |\|\ \"|\ \ |\|\ \_____| __________
: > _____ \|\|\ \ \|\|\ \ \ \ \|\|\ ___\ \ Colour /
: >|\ \ \|\|\ \ \|\|\ \__\ \ \|\|\ \ | \ This /
: >|\|\ ```````` \ \|\|\ \ \|\|\ ````````\ \Pert/
: > \|\|""""""""""""""""| \|\|""""""""""""| \|\|""""""""""| \h!/
: > \|________________| \|____________| \|__________| \/
: >
: Figure 1.12: the finished product
a very good lesson, Michael, but I think you missed that in adding the
horizontal bar to indicate depth in the interior of the O, you also need to
shade the left surface the same way all the other left surfaces are shaded,
thus:
_____ _____________ ____________
|\ \ |\ \ |\ _______\
|\|\ \ |\|\ \"|\ \ |\|\ \_____| __________
_____ \|\|\ \ \|\|\ \\|\ \ \|\|\ ___\ \ Colour /
|\ \ \|\|\ \ \|\|\ \\|\ \ \|\|\ \ | \ This /
|\|\ ```````` \ \|\|\ ```` \ \|\|\ ````````\ \Pert/
\|\|""""""""""""""""| \|\|""""""""""""| \|\|""""""""""| \h!/
\|________________| \|____________| \|__________| \/
unfortunately, the double \s look kinda goofy, but if we were to resize the O:
_____ ______________ ____________
|\ \ |\ \ |\ _______\
|\|\ \ |\|\ \""|\ \ |\|\ \_____| __________
_____ \|\|\ \ \|\|\ \|\|\ \ \|\|\ ___\ \ Colour /
|\ \ \|\|\ \ \|\|\ \|\|\ \ \|\|\ \ | \ This /
|\|\ ```````` \ \|\|\ ````` \ \|\|\ ````````\ \Pert/
\|\|""""""""""""""""| \|\|"""""""""""""| \|\|""""""""""| \h!/
\|________________| \|_____________| \|__________| \/
finally, I think if you advocate consistent use of "" as a top-front border
for the interior of the O, you should do the same for the E.
then I also regularized the spacing (the E was one space further away from the
O than the O was from the J) and made the letters the same width because I
_hate_ proportional fonts
_____ _________________ _________________
|\ \ |\ \ |\ \
|\|\ \ |\|\ \"""""|\ \ |\|\ \""""""""""| __________
_____ \|\|\ \ \|\|\ \___|\|\ \ \|\|\ ``````````\ \ IHNWIJ /
|\ \ \|\|\ \ \|\|\ \ \|\|\ \ \|\|\ \"""""""| \ LTS" /
|\|\ ```````` \ \|\|\ ```````` \ \|\|\ `````````````\ \Pert/
\|\|""""""""""""""""| \|\|""""""""""""""""| \|\|""""""""""""""""| \h"/
\|________________| \|________________| \|________________| \/
but maybe that's just me...