|
 |
Marc Schimmler wrote:
> This is just what happened to me (so I haven't voted yet)! This also the
> reason why I asked for some good advice on voting. To my disappointment
> nobody answered. Maybe the question was stupid.
It wasn't a stupid question. I suspect that no one answered you because
there is no real consensus on "how to vote." Different voters have
different priorities, different esthetic sense and different degrees of
background knowledge in 3d graphics. The hope is that these varied
viewpoints will even out and produce a good set of winners. Usually,
they do. For whatever good it will do, I will outline some of my
methods and criteria. First, I quickly go through the images with a
slideshow program. I do not study them in detail at this point, I am
just looking for common themes, the overall quality level and things
like that. One thing this does is, if many people come up with the same
interpretation, I know it before I start detailed judging so that the
knowledge affects all renditions of that particular idea equally. I
then load them into Winvote, read each text file, and look more
carefully at each image with the textfile info in mind. This helps me
to spot clever technical tricks and artistic decisions and judge their
success. It also ensures that I have all needed info to recognise and
judge the idea of the image and how well it presents it. In some cases,
I look at the zip file, e.g. to determine whether a particular object is
a csg, a custom mesh or a canned object, so that I can give due credit
for technical skill. This is seldom needed. I then go through, vote
the images, and comment them as I vote, flipping frequently between the
image, the text and my comments. The vote helps me see just how well I
like the image and which aspects to comment on. I then read my comments
again in order to decide whether the vote was, in retrospect, correct
and fix it if needed. When I am done voting and commenting, I go
through all of the votes one last time, looking for ones that no longer
seem right (I do this over a period of several days and I want to be
sure that, say, a bad mood one day doesn't make some scores too low.)
Then I compile the whole mess and send it in. For specific voting
criteria, I start at 10 and add or subtract as needed. The scores I
assign tend to be low compared to the averages. I seldom go over 15 and
the highest I have assigned to date is 18. I call that "leaving room
for Picasso." This, of course, hurts no one as I do it to everybody. I
assign a preliminary score in a category based on my overall impression
of the image in that respect. I then add or subtract if needed for
special circumstances. In artistic, I look for either overall beauty or
artistic ugliness, composition (placement of objects in relation to the
viewpoint, use of dynamic asymmetry, etc.), use of color, lighting,
appropriateness of textures, realism (if the image tried to be
realistic), selectivity (does everything in the image contribute to
it?), good use of detail, and stylistic coherence (is one key object
very realistic, while the rest are cartoonish?). In technical, I look
for original and clever techniques, whether the textures are well
designed, whether the artist is pushing the limits of the renderer, how
well the artist overcomes the limits of the renderer, how intelligently
the artist exploited the strengths of the renderer, how well made custom
meshes, csg's and height fields are, whether the artist created new
tools for the image, whether the textures and colors are custom created
(and how well) and things like that. How well this can be judged
depends on your knowledge about the various rendering programs. If
something is flashy, but is routine for the renderer (e.g. Bryce
mountainous landscapes), it is worth less technically to me than the
same thing done in a renderer not specialized for that purpose. I also
penalize technical rules violations (such as illegal postprocessing)
here if I think it is warranted. In concept, I look for whether the
image is on topic, how original and clever the idea is, whether the
image actually shows the alleged idea, how strained the connection to
the topic is, etc. If the connection to the topic is unexpected, but
good it adds points. If it is a stretch or an irrelevant detail thrown
in to "qualify" an otherwise irrelevant sci fi, fantasy, car, tv, comic,
etc fan image it loses lots of points. I penalize here for rules
violations such as entering an image made before the round began or
entering an image with no actual rendering involved. The main thing I
look for in concept (and it tends to spill over into artistic) is: "Does
the image say something" Does it tell a story, make a philosophical
point or illustrate a meaningful event. A random building, for
example, is just a technical exercise. It probably isn't saying
anything. The great pyramid being constructed, a rotting hulk in an
otherwise vibrant downtown area, or an old barn with encroaching condos
and strip malls in the background are saying something. This is hard to
explain, but I hope you see what I mean. If an image is a blatant
violation of any rule, I lower all category scores as well. When I
comment, I try to find something positive to say about each image, but
if I see nothing good ... so be it. I confine negative comments to
aspects of the image itself (I might say "Your textures are much too
simplistic," but I won't say "You're a lousy artist.") In scoring, I
try to judge every image by the same criteria, but in commenting, I try
to go easy on people new to rendering, and concentrate on general
suggestions for improvement, while I am much more likely to nitpick work
by established masters. I could go on at some length, but I think you
get the idea. Remember that none of this is obligatory or set in stone
and other people have completely different criteria and methods. This
is just how I do it.
Jerry Anning
cle### [at] dhol com
Post a reply to this message
|
|
