As a traditionally trained sculptor, working with Zbrush on a sculptural project throws up many problems. Some of these are related to the disparity in expectations of the digital medium relative to material. Other are caused by the inherent qualities in the digital tool.
The use of software tools is inherently risky e.g. an error in a file can lead to its catastrophic loss. ZBrush itself is not entirely stable. It crashes quite regularly. The ways of dealing with this risk are to make regular iterative saves of your work. this might sound like a simple solution but it is not without consequences. ZBrush is capable of saving files that it cannot subsequently open. This can happen as a result of the size of the file exceeding 4 gigabytes in memory, or as a result of a faulty mesh which can happen without warning. So it is necessary to to check the file once it is saved. This whole saving process can take a couple of minutes, which means your attention is taken away from the task of sculpting.
Breaking away from the task is not inevitably a bad thing, it only becomes so when you have to do too often. The frequency of saving is related to how well you trust the software and in the case of ZBrush I am having to save every 10 mins. This is actually intrusive and has a fragmentary effect on the attention I can give to the sculpting tasks. even with this frequency I still regularly lose work and have to repeat it often.
This is a critical problem, because sculpting a surface on screen is highly dependent on building a working memory of the forms in short term memory. when you make adjustments to the sculpt they are in relation to a mental model you have constructed from rotating the mesh on screen. This is highly vulnerable to distraction and when an interruption takes your attention away you have to start at the beginning and create a new mental model.
Apart from the basic problem of stability, ZBrush has several behaviours that distract attention from the task. The worst is probably the mis-interpretation of pen gestures. The authors say that they have deliberately build it a set of behaviours that read certain gestures of the graphics tablet pen as navigational cues. e.g. moving it side to side with the shift key held down produces a jump zoom of the view of the mesh you are working on. up and down produces a different jump and if part of a circular gesture is detected it rotates the mesh arbitrarily. The result when working at speed is regular arbitrary repositioning of the mesh, which means you have to reposition it in order to work on it. this wastes a huge amount of time. and create high levels of frustration that decreases the overall quality of formal decision making. Pixologic maintain that it is intentional but I fail to see any value in this behaviour when the normal navigation in ZBrush is so effective in all other ways.
It is a bad enough problem cause me to switch to a competing package, Mudbox when I have completed this project.
Evaluating the digital process relative to working with material. The most conspicuous problem is that of catastrophic loss of work that has already been done. However a more important problem in terms of it’s effect on the artistic quality, is the difficulty of creating the object in mind from only visual cues. With clay there is a kinaesthetic memory being produced in conjunction with the visual, and the two together create a more persistant mental model of the sculptured form. With on screen digital work the difficulty is exacerbated by the fact that the visual cues are far from accurate. The lighting in ZBrush is shader based, which mean that shadows are calculated as approximate adjustments to the colour on the surface of the mesh. They are not cast shadows. The approximation works well froma distance but is highly confusing to the eye when you zoom in. You simply cannot make confident evaluations of relative volumes or the curvature on the surface. This means that you are forced to rely on checking and rechecking the silhouette in order to build certainty about the shape. Again, this is fine at a distance but fails in close work.
There are traditional sculptors, notably Auguste Rodin who use methods that build form based on looking at silhouettes, so it is not impossible to produce good sculptural work by referring to the profiles of the form. it is however a major limitatation in terms of using a software which does not give the ability to directly edit the parts of the mesh that are perpendicular to the camera view. It would give huge improvement to the speed of working if it were possible be to isolate the parts of the mesh that are close the plane that is perpendicular to the camera view. It would seem an easy thing to achieve for a programmer, however Pixologic have stated that they see no use for the function.
I notice in a lot of digital sculpture work is lacking in depth of form, I suspect that the lighting problem is the cause. This is an issue I hope will be solved by Real Time Ray tracing in the future. The other problem of creating accurate mental models
may not be solvable. In theory, haptic feedback interface methods should help in this respect, but in reality it is an impractical solution and does not bring enough improvement.