The Perfect Artist is a video portrait installation that explores creativity: the people who create and what it means to be creative. The project was exhibited as part of Present Tense, at the National Portrait Gallery from May to August 2010.
The project webpage can be found at: perfectartist.net
Artists, arts workers and administrators were invited from around Australia to have their portraits recorded. The shoots were conducted in Melbourne, Perth, Sydney and the Northern Territory.
These portraits form a unique and colourful database of 84 individual creative people ‘performing themselves’. In total there is around 50+ hours of footage, edited into 1400 clips and databased according to subject matter, emotional tone and performative criteria.
The portrait process was designed to enable a live performance involving a series of improvised vignettes. Whilst participants were provided with an overall structure, there was no rehearsal, script or traditional direction. Participants choose what to wear, what to bring and what to say.
Participants were invited through host organisations, social networking sites and personal referral. The invitation process asked for each participant to bring along a series of items from their personal and professional lives: something from their bathroom and kitchen, a book, a song, something from there work practice, etc. These formed the basis of the categories we shot, with each participant improvising an action and/or speaking about what they had brought.
Reflecting the live performance of the participants, the final installation consists of a computer ‘performing live’ in the exhibition space – generatively sorting, selecting and composing the video clips in real-time.
The software is written using openFrameworks in c/c++, and uses several strategies to decide on what mix of categories, cities, genders, lengths of clips, amount of audio and movement to show. I am using a method called "constraint resolution" to both choose which clips to show, and how to arrange them visually and over time (for the nerds out there I am using a library called Gecode).
Basically (?), the program first analyses every clip visually and aurally (non-realtime), determining the boundaries of where and how much people moved around, and when they started and finished talking. Then the program sorts all the clips into a series of subsets - so for example every clip about domesticity goes into one set; every clip from Darwin into another; every clip with a high level of performance into yet another, etc.
This results in a large number of "overlapping" sets, from which the program can create a ~10 minute long vignette by finding the "intersection" of, let's say, 7 clips that have at least 2 clips where people are holding signs, a total of 6 clips (including the 2 signs) from the category Hygiene, 1 clip about Making and Creating, with at least 1 person from every city and as many clips as possible with an emotionally positive tone.
The resulting set of 7 clips are then given to a second part of the program which arranges them across a "timeline" with no visual overlap, and so that only one person is ever speaking at any one time. It also tries to minimize the total length of the vignette, and maximize the distribution of where people appear on the screen.
As each clip is played, it is taken out of the original sets, so that each day, and every vignette contains different people. When the program can no longer "solve" - or produce a vignette - because there are not enough clips, the sets are re-populated with everyone again, and the cycle begins over.
So much for "basically". It does a bunch of other things too, like select and play introductory music for each vignette based on 8 compositions (each with a possible 5 parts) by Kelly Ryall, show didactic information about who is on screen, spatialize the sound according to where people are standing and generally make everything look and sound spiffy :-)
The Perfect Artist was made possible with support from the following perfect organisations: Arts Victoria, Aphids, PVI Collective, Performance Space, PACT, Head Quarters, Bill & George, Tiny Stadiums and the National Portrait Gallery.
Loading more stuff…
Hmm…it looks like things are taking a while to load. Try again?