3d visualisation
- There is perhaps the misconception that architectural visualisation
is something of a dark art when in fact it is not; as a result this
article intends to inform of the ways and means that the discipline
brings two-dimensional drawings to life.
The process of creating a
finished 3D rendering is much the same as its reason for being in that
it is a communication-driven dialogue between all parties who have an
interest in the final product. Sometimes it is left in the hands of a
single body and at others it is a cross-discipline exercise.
The
very first plank with which to create a successful rendered image is a
drawing. More often that not these take the form of digital files from
applications such as AutoCAD or MicroStation and contain a vast amount
of technical data which is of little relevance to the visualiser; the
first task therefore is to remove the information that will have no
influence on the final model(s), thus making the drawing a simplified
version of its original self.
Before we move on to the next
stages it should be clarified that whilst digital files are commonplace
in this day and age, it is by no means a statutory requirement; indeed, a
visualiser should be able to work from hand-drawn plans using a scale
ruler or even sketches supplied by the client, the latter being where
important aspects of dialogue takes place as it is through this
conversation and interrogation that specific and important details are
revealed.
Returning to the assumed path of supplied digital file
and its simplification, the next step is to bring the file into the
modelling software. Most commonly this will be 3D Studio Max (also
called 3DS Max, Max and Studio Max) but other applications such as
Cinema4D and Maya proliferate.
Tangential to this, it should be
noted that clients are also supplying visualisers with digital models to
supplement or replacement a drawing. Products such as Google SketchUp
have simplified this process and it is now the option of choice for a
number of practices and studios. Perhaps the only drawback of this is an
'unclean' model whereby what appears acceptable to the naked eye is in
fact a mass of broken and co-planar (two separate planes of geometry
sharing the same three-dimensional space) faces which require tidying up
or, in a worst case scenario, complete remodelling which sadly obviates
the usefulness of their original submission.
Returning to
bringing the file into the modelling software, if the visualiser is
using 3DS Max then the file will often be imported and then placed in
its own layer which is frozen so as to protect it from inadvertent
modification or editing. Thereafter the visualiser will begin to create
all the elements required to complete the shot, perhaps by tracing the
drawing and extruding to the desired thickness or by starting with a
flat plane and manipulating it to suit - modelling techniques are myriad
it has to be said.
When the basic model is completed - the
principals described above applying to all elements of the scene -
cameras and lighting will be added and the project will begin to become
rationalised. This process will involve drafts of the image rendered at
low resolution and also at lower quality than will be used for the final
image; at this stage the most important aspects are to resolve the
camera and lighting positions, nothing more.
This development
brings us quite neatly to the rendering engine: a rendering engine is
the software which is used to create the final visualisation. VRay is
one, if not the most, common product although others such as Mental Ray,
Brazil and Maxwell are used, along with several others - each has its
own advantages and niche and for the purpose of this article we will be
using VRay as our example.
The rendering engine uses mathematical
algorithms to calculate how light and materials interact with one
another. The visualiser can have a direct effect on the quality of the
calculation by way of increasing or decreasing a number of settings
related to how seriously the renderer will interrogate the scene - in
short, lower settings give a broad-brush overview which will not pick up
detail whilst higher settings will use a much finer brush to pick up
each and every nuance, albeit at the cost of rendering time.
Going
back to the set-up of the scene in preparation for the next step, this
is where one of the most important stages of dialogue occurs. Prior to
this point the visualiser has been able to work relatively remotely, on
the assumption that the package of drawings and/or models is suitably
comprehensive. However, now is the point where the client or clients
become involved and comment on their vision for the finished shot and
how that relates to the sample(s) provided so far. It is not a one-way
street though - the visualiser will offer comment on what is achievable
and the ramifications of pressing too far with a specific demand for
instance. This segment of the process may take a number of iterations
until a mutually-agreeable position is found but it is undoubtedly a
vital part of the project.
Once the camera and lighting have been
fixed the visualiser will then progress to texturing the model.
Texturing, put simply, is the application of image files to specific
areas which imitate the actual physical finish. These files may be based
on resource photographs of the actual material or be purely digital
files which have been created using a reference. More often than not it
is actually a combination of the two; this is because a reference
photograph will 'tile' - the same detail repeated over and over again
throughout the scene and therefore belying its origins - and it is one
of the visualiser's greatest tasks to eradicate that occurrence in order
to make the texture appear seamless.
As we have touched on,
textures can comprise many sources, be that a direct referential
photograph, an entirely digital interpretation or a combination of the
two. There are a number of different ways to produce textures, with the
three mentioned above often carried out in an image editing application
such as Photoshop, whilst there are also specific applications to create
textures such as large-scale versions of bricks. The visualiser will
always use the most appropriate, accurate and efficient tool for the
job.
Once the texturing has been completed, low-quality
renderings will be submitted for final comment. It is at this stage that
opinions on the look and feel of materials are given and acted upon and
it is not uncommon to need to make iterative changes to the materials
within the scene in order to make them as physically real as possible.
At other times some are replaced entirely as they do not live up to the
client's expectations whilst others may be modified in post-production
(the stage after rendering). The visualiser will advise on the most
appropriate course of action, often based on time-effectiveness as it
can be easier to manipulate a tonal variation inside the rendered image
than spend several hours attempting variants upon a theme in the scene
itself.
Once the client is satisfied with the sample visual it
will be sent to render. This may be on the visualiser's own machine or
using a network (also called a render farm) of highly-specified slave
devices. This latter option is usually only available to larger studios
but there is a growing proliferation of vendors who effectively hire out
banks of machines for use and charge by the processor and by the hour,
this being an invaluable resource to the lone freelancer with a huge job
to complete.
3d visualisation
As soon as the image is
rendered, it can be taken on to its final stage before submission, that
being post-production. Utilising image editing software such as the
aforementioned Photoshop, the visualiser will proceed to add to the base
rendered image with elements such as a suitable background image and
sky, vignettes, camera blooms and a host of other edits with which to
improve the final visual and match the client's brief at which point the
final image is sent for sign-off and the job brought to a successful
conclusion.