[Freeware]
Introduction to the creation of photorealistic synthetic images
1.
Imitation of natural
perspective in photorealistic tridimensional representations.
The gretest
artists of the italian Renaissance were among the first ones to focus on the use
of special tecniques that could be deceptive for humans eyes, by simulating
depth on bidimensional surfaces. This illusion is based on a whole of
geometrical knowledges that dates far back to Greek civilizations.
Renaissance artist learned to turn these knowledges into works of arts which were characterized by an impressive realism. Up to the artists had been using art to represent ideas, Renaissance artist besides began to consider art as a particular device which could be used in order to simulate reality. That marked a huge fundamental changing.
By this time, more than a century of cinema and photography has deeply explored the
tridimensional space representation on straight surfaces. Movies mixing up
animation and tridimensional images with photographic and cinematographic
traditional tecniques led our attention towards a subject whose aim is to
accuraty shape phisicaland optic proprieties of the real word.
Again, art and mathematics joined together to charm and guide the observer into the
unreal but sinthetic imaginary, allowing human mind to be able of a fantastic
creation of the immaterial.
Photorealis
describes images created by computers using phisical and optic proprieties of
the real world, mathematically translated, so that an observer find himself
deceived at the point that he experiences a hard time in telling an artificial
image from a photograph.
The ability of the computer operator and of the computer itself in simulating photographics images particularly striknes all those persons thinking to computer just as aseptic instruments, exclusivelly mathematics, without any regard to their real potentiality on this topic; they can be considered as tools that could substituite brushes and palette of past artists.
Anyone
having ever found himself in drawing a tridimensional form, certainly
experienced difficulty while creating depht illusion on a stright surface. By
the way, tridimensional space concept is likely to be so abstract that even
children don’t have an innate comprehension of concepts like depth and volume
themselves.
Learning of objects shaping in the
space is, in a way, so important that basic principles are often forgotten; even
if not so much time haspassed since artits began to appreciate and exploit the
potentialities.
Passage was
almost istant : panel on the left shows a Cimabue work
(1240-1302 , 1280 ); panel on the right a Giotto work ( 1266-1336, 1310 and than
30 years later ). We can see them placed together on the same wall at the Uffizi
Gallery of Florence; though separated by a huge conceptual difference.
Cimabue was interested in
tridimensional representation ( one can easily deduce that observing the form of
the throne and arches below ), nevertheless it hasn’t obtained a realistic
simulation: in fact, in a photograph it would’n be possibile appreciate, at
the same time, the internal wall of all the three arches. Therefore it comes out
the impossibility in finding a correct point where we could place our
observation point; so, the painting itself, in its entire forms, is likely to
seem unrealistic if to a well used to tridimensional eye observation.
The 'Calice' of Paolo Uccello (1465) Computerized
elaboration of the 'Calice' Utah Teapot (object of reference for comparison of the qualities of the algorithms)
·
Wireframe
·
Flat ·
Gouround The power of light is shed over the
entire polygonal, allowing to fade side edges but, still, surface remains
not continuous and little realistic if we don’t make use of other
corretive algorithms. Calculation time is quite fast.
Commonly used standard algorithms are 3 : Each pixel of the object is calculated
on the basis of the whole light involving it ( reflected one included ) by
tracing luminous ray path from a point of the screen to all the objects of the
scene and, afterwards, to the luminous source. Reflections, refractions
and shades are moulded. All the object iterations, lights and
screen are defined, including luminous rays not reaching the screen. This
is the most precise modality but a so high calculation time ( few ours )
restricts its usage just to great systems and for particular necessities. with once of enough
rendering ( to the endless one), the final render will always converge toward
the correct solution without the introduction of artifices. The other motors of
rendering or other methods of interpolation always produce ‘biased renders, they
cannot independently guarantee consequently the convergence toward the correct
solution from the planned computational time. This algorithm is what produces
the most elevated quality and is currently used by the most greater
producers of CAD 3D professional (3dsMax, Viz, Maya, Lightwave, Rhino,
Solidworks, ArchiCAD, Cinema 4D, Blender, etc.). Photorealistic effects of reflection and shading of an algorithm of
ray-tracing.
Which are the real photo and the images photorealistic created with an algorithm
of radiosity ? TILECAD foresees the use of the
following computational algorithms for the creation of images 3D:
-
OpenGL ( MESA )
: algorithm phong for the
immediate calculation enriched with effects of shades
-
POV-Ray
:
procedures for the calculation in
ray-tracing
-
MEGAPOV
:
enrichment for the calculation in
radiosity
-
INDIGO
: procedures ‘unbiased
rendering
Each
of these technologies has his specificities and characteristics and their use it
is tightly in operation of the necessities of the consumer. we don't enter the
complexity mathematics-physics to understand the conceptual differences among
the single procedures but we appraise rather some parameters that qualify the synthetic image final. The computational time (consequent
also of the computational abilities of the computer system) it is certainly a
parameter to consider with attention: This diagram allows to
appraise that having available a greater computational time they are always
gotten greater quality to condition to change the typology of rendering.
Opposite with low availabilities of time it is not worthwhile to use algorithms
that require for them nature of strong potentialities of elaboration.The consumer has to
appraise in advance the ‘s resources that you intends to use and the
availability of attended before getting the result. In operation of this
decision can activate the consequent procedure. He could for instance reassume
this way: low computational abilities and immediate application:
Phong averages computational ability and application within some
about ten minutes: Ray-tracing and/or Radiosity tall computational abilities and application without
limits of time: ‘Unbiased '
Example of render with algorithm ‘unbiased' in an inside: to get the
effect of external solar light that penetrates from the window and it radiates
him in the environment and on the objects, reflecting itself and refracting
itself, necessary 12 hours of elaboration have been with a workstation of
averages ability. The program 2D/ 3D of plain’s composition TILECAD laying, you in his releases of distribution
customize, it use Flat algorithms
for the visualization in real time 3D, algorithms of Gouround and Phong, altered
and adulterate, for the visualizations to average and tall definition.
Foresee systems hardware of elevated power. Creation of preliminary images to lower part number of
pixels.
Uses of the algorithm of alone radiosity like conclusive
process. Accurate control of the points light.
Attention to the effects of perspective that they could be
deforming. Uses procedural respect to the maps of
bits of the texture. Threshold the complexity of the scene especially on the
backgrounds.
The 3D-stereoscopic images are known from beyond 100 years and they am corporate beginning from 2 different superposed images in such way that each eye sees only ‘his’ image. Because of the finite resolution of the monitor, the best manner for realize 3D-images on a computer it is that said of the ANAGLIPH: an image comes ‘encoded’ in red, the other in blue or green, after of that it is proceeded with the overlap. For see (‘decipher’) the images are necessary of the special glasses with red and blue lenses (or red and green). Such glasses, realized in plastics or carton, they are very economic and they could be acquired for few money. TILECAD create the anagliph images
with a specific option to the inland of the program. An example of this
elaboration is the inferior image. For preview it you bring the glasses to the
eyes (the red lens must be to the left hand) and gather on the image. After a long codifies and test has been produced completely from OMNI DATA a new
algorithm of
visualization 3D for
TILECAD© of
the type ' not-fotorealistic render - Sketch'
and
that he already adds to the 7 existing typologies.
Totally in automatic and immediate way, the procedure
transforms the visualization 3D in a fac-similar of sketch ' hand drafting' /
comic strip, in tonality of grey with sbording of the straight lines, loss of
the colors on you furnish to him and outline of the images in transparency.
The result is saving in one file or printable for the following
elaborations that the user you can effect using very quickly or computerized
programs of color-draw or the usual tools artistic manuals (ink, pencils, pieces
of chalk, tempers up to the use of the aereograf).
To drive the algorithm and to get the most consistent image to necessities, the
user can choose between 3 typologies of impression of writing: also
intervening on the choice of visualization of you furnish 3D to him and on the
layings of the tiles. Some algorithms studied on purpose for this new
visualization are also proposed in the typology ' Phong' to low quality.
To conclusion it is gotten totally practically an
indistinguishable sketch from one appreciated production of impression manual
and artistic losing so the computer connotation.
Within the thirty years
passed from Cimabue fron Giotto works, artists has discovered a new powerfull
method in order to guide the observer while comparing different paintings.
Giotto painting is seen with a more relaxing sensation, thogh it isn’t perfect
yet.
It is necessary to wait till Renaissance
(1450-1500) to gain mathematic codification of perspective and its correct
graphic application.
Piero della Francesca (
Pala Urbinate – 1472 ), Giorgione, Leonardo da Vinci, Correggio and later on
all subsequent artists considered perspective and tridimensional draw tecniques
as a philosophical thought, where order of things is likely to coincide within
the mathematics of the form and, therefore, an extract measure of the real
world.
Escape point is clearly determinated, tridimensional
solidities are conspicous, background is a lettle hazy, foreground is brighter
and more contrasted, volumes and dimensionals are correct with regards of depth.
These statements convince human eye to imagine a depth
right where it can’t exist and the artistic collocation light up natural world
to the same level of spiritual world typical of the religious content. Therefore,
we can say : a perfect mathematic coordination between divine and earthy.
Since technology takes place between artist and his work,
it would be likely that one easily thinks that art is turning into something too
mechanic, without any sort of inspiration. However, the Renaissence ‘happy
matrimony’ of science, technology and art contradicts this hypothesis.
To reach a clearer comprehension of the matter above an
example follows: the well-known ‘Calice’ from Paolo Uccello ( 1465 ). It
results composed of segments when subdividing the surface into rectangular
geometric forms. These last seem to us definitely similar to polygons abtained
with nowadays computerized images ( wireframe ); it is quite striking
thinking that this happened six centuries before computer elaborations.
1.
Creation of synthetic images with the computer graphics.
At the beginning, computerized tridimensional images were
created by means of mainframe computer using direct mathematic formulations.
Currently, a lot of softwares allow us to manage complex graphics by means of a ‘frendly interface’, that simplifies access to several functionalities of
program and makes objects elaborations, scene and lights much more openly
intuitive.
Creations process can be schematized with the following 4
phases :
1.1
Planning
This is the creation of
storyboard . One has to define the ideation of the scene just like an artist creates
with a preliminary sketch, the work he is going to realize. In computerized
elaborations storyboard is mainly usefull while planning the complexity and to
obtain an overall approval before the productive process gets on the way.
1.2
Shaping
Shaping is the process of
creation of objects geometry and assignation of surfaces to objects themselves. At
the moment, many different tecnicques allow us to accelerate this kind of
procedure that, otherwise, would be very complex: extrusion from 2D,
interpolation, sculpting from 3D, procedural methodology, etc. with care of the
program we are going to use. Therefore, 3D objects resulted from the work above
can be composed of polygons in the space similar to an iron cage, or
bidimensional surfaces in the space ( Bézier’s, spline,etc. ) or solid
massive compounds i.e. with barycenter of masses.
In any case, a pixel surface is
applied on these objects, using a reference map in different ways: assignation
of a pictorial texture (winding), assignation of procedurale textures (fractal
algorithms) or by simple shading ( color plus light ). 1.3 Scene
The preparation of the scene is
carried on with the tipicall way adopted for theater productions, i.e. with
the collocation of objects within definite space, camera, equipping of lights
and, eventually, animation actions ( frames ). The planning out of the scene can
also place with the simoultaneous partecipation of an aesthetic evaluation of
composition that will be, afterwards, the final resul on this elaboration
1.4
Rendering
It is the operation for automatic conversion of objects and their
attributes turned into one or more fit-to-print and file syntetic images. During this work-period, computerized systems make plenty of calculation and even the
most sofisticated computer potentiality is put to a hard test. This is
definitely the most complicated computer elaboration man has ever conceived and since
just few years we are able to create synthetic images of great
sophistication in moderately brief periods of time
1.
Algorithms for 3D visualization.
Light and shandings are the main parameters that define
higher or lowel of photoreality in a scene. There are several methods for
shading; they can be condensed as follows:
Objects have diaphanous surfaces that
allow us to see hidden lines and points. This method is exclusively
geometric , not photorealistic, and permits a real-time visualization while moulding
the scene.
A single level of illumination is
assigned to the entire polygonal surface of objects, so that single faces (
mesch ) show a uniform dull color and differents each others. This
definition doesn’t produce so natural results, but it requires a little time
for elaboration. It is mainly used for preview visualization.
This definition produce a surely more realistic result. Each pixel of the
surface is calculated with regard of the luminous source by means of reflection
effects. The number of necessary calculations is much greater if compared to
previous method and it increases if the image resolution increase.
·
Scanline
Where the color of a point of an
object is determined through its relationship with the observer and with the
luminous sources of the scene itself. Reflections, refractions and shades
are not moulded. ·
Ray-tracing
· Radiosity
·
'unbiased rendering'
TILECAD software links.
POV-Ray,
MEGAPOV e
INDIGO
it
foresees be algorithms of ray-tracing that of radiosity for the creation of
synthetic images, as like other software of visulisation 3D.
I use it practically it of all the algorithms of
visualization allows to the consumer of obtain levels of photorealism
to his discretion: from that simpler and foreseeable up to that more
adulterate but with times of very tall elaboration.
Of succession see the general structure that we foresee for
an I use optimal of these tools:
That the structure of the scene 3D has given birth to
automatically from TILECAD, I for him use of software with algorithms of
ray-tracing / radiosity, the consumer must stop particularly on the correction
of the lights/ shades and on the procedural texture that such software allows.
In fact also if TILECAD inserts in the preparation of the
line of conversion all the graphic information and you of visualization
that has been use that for the visualization to his inlan it is not said
that it am sufficient for obtain synthetic optimal images, sight the
modification of the used algorithms.
It is advised to read the documentation that we allege and
of experiment directly with the present examples. Additionally seen the long
times of calculus and the informatic weight of these modes consults:
3D-Stereoscopic images (Anagliph)
6. Visualization 'Sketch - no photorealistic render' (hand drafting / cartoons)
