[-empyre-] visualization as the new language of theory
Thomas LaMarre, Prof.
thomas.lamarre at mcgill.ca
Sun Feb 7 01:33:02 EST 2010
Thanks for such cogent and insightful presentation of the question of continuous variation in digitally produced images. This is one of the really interesting developments in animation and cinema, and as you've noted elsewhere (if I am not mistaken), it is one that entirely shifts the relation between cinema and animation and even between full and limited animation, in a way that makes animation generally an important paradigm for understanding new media - due the the greater emphasis on what happens to a broad range of elements within the image or in the frame, which leads to transformations in movement between images (at the same time that it is harder to speak of a single discrete space between images in the manner of celluloid filming with snap by snap footage).
Two things interest me about how the continuous variation a number of elements in the image impacts designers who are working on films, series and OAVs that are still called animation.
First, even though today there is greater control over the variation of a number of elements with digital computers, there are techno-aesthetic precedents within certain lineages of limited animation, which were worked out between interactors (otaku fans) and producers (who were themselves otaku fans), from the late 1960s, but especially in the 1970s and 1980s. For instance, with the introduction of the VCR, 'viewers' transformed into co-producers or co-operators because they began to watch these animations in terms of variations of elements within the image by replaying the animations over and over again, noting how the variation of elements (changes in colour and line) could be 'read' independently of any overall narrative or directorial vision (usually not much in evidence). In fact, this sort of interaction began to take precedence over the modes of perception and interaction that today we tend to attribute to analog media. So, with the development of digital computers, there was not a strong sense of a break but rather a transition that confirmed, reinforced, and extended what was already happening. Many animators avoided the American models for digital animation entirely, seeing them as reactionary throwbacks designed to hide or compensate for continuous variation.
I don't think it a coincidence that this happened in the context of large-volume production across a number of small heterogeneous studios in Japan, rather than in the much smaller volume large studio productions typical of North America (although there was always a great deal of interplay between these at some level). I also don't think it a coincidence that it is the same animations that enjoy the greatest transnational popularity, not to mention the greatest presence on line, and go hand in hand with micro-masses (sometimes called fan cultures or subcultures). See Tinami and Wikipedia. In other words, these transformations are worked out not merely at the level of producers making animations to consumers (classical communication/consumption model) but also at the level of socio-aesthetic formations wherein the role of consumer (already reworked socially by shifts in the relation between production and consumption) shifts toward interaction or co-production. The idea of a cooperator in emergence theory probably captures something of this interaction better than terms like co-producer, though.
Second, if we take seriously the long history of scientific thinking about, and experimentation on, materiality that begins, in the late 20th century, to draw more and more heavily on radical empirical notions of continuity and continuous variation (non-atomistic approaches), it seems to me that the term 'continuous' should not mislead us in thinking continuity or continuum in simplistic brute materialist terms. Rather, the materially and phenomenologically continuous is at once discrete and continuous. It implies a spacing that is traversed by a force or forces (to use the term generically), in a non-dualistic or interactive dualistic way. So the question of the impact of such 'constant change' might well be asked at the level of 'material spacing' and 'force.' Constant change may well become a ground that is at once acknowledged and ignored at the level of media competency and literacy. Nonetheless, we're still dealing with forces, and thus a mobile active ground and not a foundational moment that is itself unchanging. (Analogously, even if there is greater emphasis on what is in the frame, there are still questions about saturation and about outside of the frame, which cannot be kept forever outside.) Otherwise, it is a model that cannot accept disruption or disjunction in any guise, in defiance of what we actually think to be true scientifically of matter and materiality.
It is to address such issues that I proposed that we consider animations not so much as objects (or conversely ideas) but in terms of a machine in Guattari's sense. The emergence of an animetic machine - a materially specific spacing based on technical assemblages that channel a force - can be demonstrated empirically. And interestingly enough, this machine (again not an object or a form/idea) persists throughout the introduction of new techniques and expressive machines precisely because it relies on its spacing and thus on a harnessing of force, to fold expressive machines into it (and to explicate new ones). Thus continuous variation can be folded into the animetic machine, as I discussed above (issue 1).
Of course the question of whether the 'continuous variation machine' (if such it is) will generate a radical inversion of these relations can be posed, but it may have to posed at the level of forces rather than objects or brute materiality. I can't help thinking that cultural analytics are in some sense designed to do precisely that, to generate such an inversion. I don't think that they are neutral techniques that will allow us to get at the truth, even if we can get at a truth or a set of truths with them.
Without experimenting with them myself, I cannot say whether they go beyond the scientifically discredited 'classical empiricism' based on one-to-one correspondences. But with an archive of well over 100,000 hours of Japanese animations (largely from the 1900s to the present, earlier materials are less well represented), I am rather curious to see what might come of a cultural analytics of them.
What is also interesting about such developments is that the 'animetic
On 05/02/10 1:48 PM, "Lev Manovich" <manovich.lev at gmail.com> wrote:
Greetings, and apologies for not being able to participate over last
couple of days
I want to add something in relation to the questions of "limited" vs.
"full animation," animation vs. cinema distinction, and movement.
In contrast to twentieth-century animation, in contemporary motion
graphics the transformations often affect the frame as a whole.
Everything inside the frame keeps changing: visual elements, their
transparency, the texture of the image, etc. In fact, if something
stays the same for a while, that is an exception rather than the norm.
Such constant change on many visual dimensions is a key feature of
motion graphics and design cinema produced today - in contrast to
contemporary feature animations or anime where many elements do not
change because of figurative nature of these works ( for instance the
colors corresponding to diff. objects normally do not change from
frame to frame.) We can connect this preference for constant change to
the particulars of software used in media design.
Digital computers allow us to represent any phenomenon or structure as
a set of variables. In the case of design and animation software, this
means that all possible forms-visual, temporal, spatial,
interactive-are similarly represented as sets of variables that can
change continuously. This new logic of form is deeply encoded in the
interfaces of software packages and the tools they provide. In 2D
animation/compositing software such as After Effects, each new object
added to the scene by a designer shows up as a long list of
variables-geometric position, color, transparency, and the like. Each
variable is immediately assigned its own channel on the timeline used
to create animation. In this way, the software literally invites the
designer to start animating various dimensions of each object in the
scene. The same logic extends to the parameters that affect the scene
as a whole, such as the virtual camera and the virtual lighting. If
you add a light to the composition, this immediately creates half a
dozen new animation channels describing the colors of the lights,
their intensity, position, orientation, and so on.
During the 1980s and 1990s, the general logic of computer
representation-that is, representing everything as variables that can
have different values-was systematically embedded throughout the
interfaces of media design software. As a result, although a
particular software application does not directly prescribe to its
users what they can and cannot do, the structure of the interface
strongly influences the designer's thinking. In the case of moving
image design, the result of having a timeline interface with multiple
channels all just waiting to be animated is that a designer usually
does animate them. If previous constraints in animation
technology-from the first optical toys in the early nineteenth century
to the standard cel animation system in the twentieth century-resulted
in an aesthetics of discrete and limited temporal changes, the
interfaces of computer animation software quickly led to a new
aesthetics: the continuous transformations of all visual elements
appearing in a frame (or of the singular image filling the frame).
This feature of contemporary motion graphics makes it particularly
suited to cultural analytics approach. It is quite easy for software
to measure any visual parameter in each frame - amount of motion, hue,
saturation, texture, shapes, lines, edges, etc - and then visualize
how each of these parameters change over time. We can then see the
patterns of change and compare them across thousands of works. (Of
course the same approach works for feature length animation or any
other forms of moving image.)
Here is an example of how such graphs look:
In this example I compare four different works:
columns (left to right):
Betty Boop cartoon, 1930s;
Bjork music video, 2005;
3D computer animation by Ann Lislegaard (Crystal World, after J. G.
video for "Go" by Common, directed by Convert/MK12/Kanye West, 2005
Row 1 (top row) = key frames
Row 2 (next from the top down) - a "slice" through a video
Row 3: visualization of changes in average (mean) brightness level
(each vertical line corresponds to the average brightness of each
Row 4: same information as a normal graph
Row 5: graph which shows frame differences (the number of pixels which
change from frame to frame)
empyre at lists.cofa.unsw.edu.au
More information about the empyre