[Digital Genesis]

I zapped this from a mailing list some time ago. I thought it might be an interesting read for those interested
-------------------------------------------

Hi Marcus -

Here's my $.02:

I don't think it's a single "coloring technique", I think that what people
call "photorealism" results from careful attention to a number of small
details which, when taken together give that impression:

1) "Consistant" anatomy (I won't say correct, or even plausible...
The anatomy of Hajima Sorayama's females is as much high fantasy
as any dragon or centaur. It is, however, consistant with our
willing suspension of disbelief.)

2) Lack of contrasting outlines (or other obtrusive rendering
stylistic conventions).

3) Consistant depth cues. (Modelling light. Thrown shadows.
foreshortening. Drawn reflections.)

4) Logical and consistant use of light.

5) Relatively uncontrived composition - Because a camera catches
*anything* within its frame, we have come to associate more
formal composition with painting and less formal composition
with photography.

To verify this, look through Elfwood and whenever you see a
picture that was drawn from a photograph, ask yourself how
you know this. It will likely be because the artist has drawn
what the camera captured of the subject - both good and bad -
whereas if the artist had composed the picture either mentally
or from multiple or life references, they would have
unconsciously stylized, abbreviated and otherwise adjusted until
the picture is "tighter" than reality.

I think "photorealism" -- and I expand the term here to encompass all
relatively "realistic" forms of rendering (as opposed to line drawing,
cubism, outline-and-fill comic art, anime, etc.) -- works differently
in the brain of the viewer than other art does.

I think the distinction has to do with the physiology of how humans
see. Contrary to some popular notions, the human brain doesn't basically
get a raster image and try to "analyze" into meaningful shapes.

The eye and optic nerve actually capture the surrounding world not as
"images" but as two distinct sets of information [lecture on the
rods and cones inside your eye deleted]. Basically, by the time the
information stream from your eye hits your cortex, it is already
separated into outlines and textures. The prevailing theory is that
the shapes are used for symbolic recognition and then textures are
reassociated for detail to complete the "seeing" process.

So, my theory is that we have a basic psychological prejudice that
labels images as "real" if most of the "fill in the blanks" texture
and detail information comes from the eye rather than having to be
invented or retrieved from memory by the brain.

If we draw simple shapes, the brain still recognises them for the most
part as what we intend them to be -- they just supply half of the
shape-and-detail stream the brain wants. Our brain can fill in the
missing texture and detail to a certain extent from memories of other
texture and detail. We are more-or-less aware that this is happening,
though, and "know" we are looking at a drawing.

But the more texture and detail cues we give that make the brain's
retrieval work less, the more we sense that an image is "real".

I think that's the essence of Fantasy and Surrealist art:
we can draw any shape that doesn't completely barf our symbolic
recognizer. As long as we provide sufficient visual texture and detail
cues, our brain will accept the implausible shape as "real".

regards,

--- Spook ---

... on an early-morning philosophical note.


Ha! More on photorealism...

I'm rolling, and can't shut up. <g>


One of the details I didn't make clear in the last post
is that I think whether we consider something "photorealistic"
is also heavily related to the level in the brain at which
we recognise it.

I mentioned that the brain receives images as a stream of shapes
and textures. Well, at some point, down in the eye and the optic
nerve we had to perform some hardware-level separation and
recognition. The eye basically had to say "That's a shape. It
goes in that stream. That's texture information. Put it in the
other stream over there."

Well,

not every shape gets recognised as a "shape" at that level. Some
things that we look at do not contain sufficient cues for the eye
to automatically rake it into the shape pile and they only get
properly categorized later on, up in the higher levels of the
brain, possibly even at a subconscious or conscious level.

Optical illusions and steganography are excellent examples of
images which don't get properly separated by the eye and require
active cooperation of the brain to correctly perceive.

So what does this have to do with a painting being photorealistic
or not?

Well, I think that we consider something "realistic" if it contains
the right kind of visual cues so that most of the primary sorting
goes on at the lowest level, right in the eye and optic perceptor.
So that by the time the brain gets the pieces it doesn't end up
doing a lot of reclassification in order to understand what it is
looking at. It is able to directly confront the image, and thus
we feel that the image is more "real". Even if we "know" that such
an image can't logically exist in the world.

That's also why bizarre color schemes can still seem realistic.
If we render something that ought to be green using blues and
oranges, but we carefully follow the rules of perspective and
shading and render the lighting accurately, our eye will happily
recognise and extract the shapes and detail. Our brain will be
able to process and understand the image without having to
reassemble the shapes until it gets something it recognises and
thus we don't question the "realism" of the image until or
unless we "notice" the color scheme at some very high conscious
level and say "Wait a minute, grass isn't purple. It's green."

On the other hand, if we faithfully color and texture another
picture, but completely muddle the visual cues - light sources,
shading, outlines, etc. The eye will separate it into shapes
and textures, but the shapes won't be ones that are immediately
recognisable to the brain. The brain will have to do considerable
symbolic processing on the shapes -- a significantly higher-level
function -- until it rearranges them into something it recognises,
at which point it can then fill in the texture and detail and
we realize what we're looking at. Still a good picture, perhaps,
but not one that we recognise as "realistic".

--- Spook ---

[CapnPyro]

argh, my brain hurts.. i read the whole thing.. but what is he trying to say? make everything as realistic as possible (anatomy, perspective, lighting, texturing) and it will look realistic? hmm... i think i already knew that... i suppose its helpful because it sort of says everything that you already subconsciesly know about drawing but havnt really thought about or formed into whole ideas. it is intersting though, how with a few quick basic shapes you can give the illusion of an entire picture, id like to learn how to do that better (sorry, no examples). oh my, look at... becoming long winded after reading that post..

-CapnPyro-

------------------
http://home1.gte.net/capnpyro