Posts Tagged ‘ambika’

UNI [THESIS] – INSTALLATION DIAGRAMS

July 22nd, 2009 | Richard Almond

Just a couple of diagrams I did for my presentation boards to help explain how my installation functioned.

UNI [THESIS] – INSTALLATION REVIEW

July 12th, 2009 | Richard Almond

This installation generally proved a useful and enjoyable exploration of my thesis topic. I have studied the metaphors of decay and memory and attempted to blur the boundaries in their definitions between the physical world and the digital. The patch began with a live camera feed being projected onto a screen. This in itself attracted a lot of interaction, being rear-projected, the feed offered a true depiction of the room and its occupants, rather than the mirror-image most are used to. Upon a visitor entering the room, a short recording was triggered. These recordings started to be fed randomly back into the projection, overlaying the live feed. It soon became confusing as to who was actually in the room and who was a previous occupant, a memory of the past. The effect was most powerful when a previous memory of the current occupant appeared. Visitors found it particularly exciting to share a space with a previous version of themselves.

Depending on the length of the installation, either every 1 hour or every 10 minutes, a new series of presets were triggered which progressively decayed the live feed. Effects were initially very subtle, simply glitching certain colours within the live feed, but moved on to drawing trails of the moving edges of an object or visitor, before eventually blending pixels into a surreal blur. Recordings were taken after the data moshing process within the patch, and so the recordings themselves were decayed. This meant that in the latter stages of the installation, the decayed live feed was overlaid with decayed recordings, leading to a doubly-decayed result. An unforeseen but enjoyable outcome was that in the latter stages, since a recording is chosen at random from the pool, there would occasionally be a memory from the very start of the installation overlaid into the feed. This recording of course had been only very subtly decayed, so visitors were periodically allowed just glimpse of reality. In the final stage, the user can no longer make out themselves as a current occupant of the space, the feed has become so decayed that they simply manipulate the pixels with their movement, pushing and pulling colour around the screen. They begin to be able to actively decay the memories of the previous occupants that they are being shown.

It soon became evident that the initially planned 6-day long version of the installation was simply too lengthly for something that is intended to be purposefully visited, and that the 1 day and 1 hour long versions were a lot more suitable. I found that visitors expected to see the results of their actions in real time. The average visit was maybe 2-3 minutes, but the shorter the overall length of the installation, the longer each visit tended to be. I feel that this is due to the immediate feedback available in the shorter versions, visitors quickly see results, they see themselves alongside previous occupants, and often even alongside a version of themselves from only a few minutes ago.

I experienced a range of technical difficulties throughout the week of experimentation. The patch itself behaved rather temperamentally, seemingly possessing a mind of its own and producing completely different results on different days. The main problem however was that of lighting, which required a very fine balance between providing enough so as to allow the distortion effect to work properly, but not too much as to wash out the projection screen. Using a series of small lamps with shields I managed to achieve an acceptable arrangement after much trial and error, but this patch would run much better on a large TV screen. The difficulty with this approach is that there is an element of poetry and depth in projecting the results, as one can choose the surface onto which to project, something showing the signs of physical decay inevitably add another level of interest. There were also problems with the storage of the recorded files, which were very memory intensive. Retrospectively, more time would have been spent finding a suitable codec to compress the file sizes without the loss of too much quality, allowing the installation to run without the need for an external hard disk. A firewire camera may have also helped speed up the general running of the patch.

This piece remains a semi-proposal, as regrettably there was not enough time to set up a user-submitted web feed of images into the installation. The idea of users being able to add to the installation from their computers or mobile phones would add yet another level of complexity to the system, and this would definitely be a consideration for future experimentation. Through this experimentation, however, I also discovered the appeal of the installation in a public place, and in particular one containing large amount of passing people. I did some testing in a corridor/room space which saw people frequently move through, and this was particularly successful. When people are unaware of the installation and not expecting to be recorded they tend to respond well. Any run of the installation which lasts more than 6 hours or so almost certainly needs to be something people passively engage with, something in a public space, rather than something which people are invited to actively visit and interact with. There is a definite possibility in using this in a venue such as a bar or club.

Experimentation with props during the installation was compelling, and this is something I would enjoy exploring further. In particular I would like to experiment with graphics, exposing words and signs during the recording process which could build up a potentially poetically random set of memories. I relate this back to “The Art of Memory” by Frances Yates, and the ancient view that there are two types of memory: that of words and that of things.

UNI [THESIS] – INSTALLATION 1 DAY PATCH 2ND RUN

July 11th, 2009 | Richard Almond

Friday was the final day of my installation, and again I ran the one-day long version down in P3 Ambika. It’s becoming clear that this patch has a life of its own, certain days will see it operate, inexplicably, completely different to the day before. Friday’s run certainly was not as successful as Thursday’s, the recording and playback of clips was suffering from an abnormally low frame-rate, which caused problems during the early part of the run. Later in the day however, the effect was not as noticeable due to the generally more decayed appearance of the projection, and in a way actually contributed to the installation. After a spot of meddling I managed to get the patch running reasonably well, and as the day was far quieter than Thursday in terms of visitors, I began to explore the potential of the patch by using various props.

There was a chair, and a sweeping brush to hand, which I began to move around the space, triggering recordings. I asked visitors to sit on the chair, and to pic up the brush, and within a couple of hours a very complex overlay was beginning to build up showing the chair and brush randomly jerking about the space. When looking at the screen it became impossible to tell where the objects actually were, which was the real object and which was pre-recorded. Visitors become very excited at the sight of themselves standing alongside a previous occupant of the space, especially if that occupant is them. They naturally try to engage with this past incarnation, mirroring their movements, attempting to grasp them.

UNI [THESIS] – INSTALLATION 1 DAY PATCH 1ST RUN

July 10th, 2009 | Richard Almond

Yesterday [Thursday] I ran the one-day long version of the patch down in P3. Most things, for the first time, ran smoothly, the camera recorded properly, the lighting levels seemed adequate and crucially there were fairly frequent visitors. The only issue was with the compression codec I had added to the recording part of the patch, which was compressing the clips to too high an extent, even on maximum setting. The codec was removed and the file storage transferred to an external hard drive, and luckily this did not seem to slow down the operation of the patch at all.

The one-day long installation lasts 6 hours, and this seems to be the optimum length of time for this patch. Results are responsive and frequent enough to maintain visitor’s interest, but the installation also runs for long enough to generate an authentic sense of passing time. As the hours go by, memories of visitors randomly appear in progressively decaying stages, and a vast range of visitors are able to contribute.

UNI [THESIS] – P3 INSTALLATION PREPARATION

July 7th, 2009 | Richard Almond

Monday was spent making the venue suitable to host my installation. Firstly, the screen was created with a simple linen sheet, which was weighted at the bottom then hung from the top on various ceiling protrusions. The screen effectively separates the space into a public section and a private section, and allows for rear-projecting, allowing visitors to fully interact with the installation. Although the projector threw a suitably-sized window onto the sheet, the expected problem of lighting levels unfortunately appeared. The data moshing patch I am using seems to operate within a rather narrow range of light quality, and tends to require fairly bright levels. Of course raising the light levels means that the resulting projection becomes washed out and visually ineffective.

Initially, I experimented with a couple of small desk lamps placed on the floor in-front of the screen. This cast a upward light which created a pleasant ambiance in the space, and I quickly discovered that the MaxMSP patch preferred only one of these lamps to be active. Further refinements saw a shield built to direct the light away from the screen and towards the visitors, and this significantly aided the patch. One factor which continued to pose problems however, was the camera itself. Unexpectedly poor quality, the camera is an external web-cam and is capturing very unnatural colour, saturation, contrast and brightness levels. All attempted tweaking so far has proved unfruitful and so an alternative may be required. The camera was positioned on a small ledge above the centre of the screen.

There was a small amount of successful experimentation done, mainly using the MacBook’s built-in iSight, the results can be seen in this Flickr set.

UNI [THESIS] – P3 AMBIKA INSTALLATION VENUE

July 5th, 2009 | Richard Almond

I’ve managed to find a suitable venue to host my installation, and a very convenient one at that, being a small storage room in P3. The space is deep in the underbelly of the University of Westminster, it has no natural light and so lends itself well to projection work, it also benefits from some degraded, decaying surfaces which allow for some interesting effects when projecting images of digital decay. I’ve done some initial experimentation in the space with projecting my MaxMSP patch, and there were a few issues thrown up which need consideration.

Mainly, there is a fine balance to be acheived in the lighting of the installation. Because there is a direct link between the camera and the projector, enough light is required to create the desired effect in the patch, but not too much as to render the resulting projection washed out.

There is also consideration to be made over the actual surface for projection. There are a number of suitable wall surfaces within the room, but due to the locations of powerpoints and the fact that part of the room is still in use as storage, it may be sensible to cordon off a section with a screen. This screen could then be used to rear-project onto, meaning that visitors may closely study what is on the screen without interfering with the projection.

Click here to see some photographs from initial projection experimentation.