Mid-exhibition Analysis of Work in Situ

Now that the exhibition has been up and running for a couple of days, I can afford to take a step back and think about the success or otherwise of the work as a whole and also of each individual vignette.

Neon Dance

This has always been the least stable of all the pieces due to the apparent inherent instability of the Carousela CV Outline plug-in. The piece must be tweaked carefully to suit environmental lighting conditions. Originally I had hoped that each of the three screens would remain blank until movement was detected in front of them at which point an outline of the viewer would be traced and the relevant rhythm element played. It seems that there is an issue with the tracing functionality becoming dormant or somehow not working if no outline is traced in  the first place. So rather than drop the piece as being unreliable I lowered the outline detection threshold so that an outline is (almost) always drawn based on ambient light. When the viewer moves in front of the webcam the shape changes to incorporate their outline. Evolving kick drum, snare and high hat patterns are linked to each of the three screens and still do sometimes re-trigger when a person walks in front of them. Normally all three tracks play continuously so the originally intended effect of movement equating to audio trigger is lost. However, the piece still looks and sounds good and generally seems to work at getting the younger viewers up on their feet.

To further improve the piece I would try to solve the dormant tracing issue and raise the threshold for outline detection. From an audio perspective I would like to modulate the group mix  a little over time, perhaps so that there appears to be some progression, which could be semi-generative.

Spark Shower

Works pretty much as intended. The threshold for creating sparks is somehow a fraction too low (and not easily adjustable)  so sparks are created even when no one is present, but it is obvious that bold movement creates many more sparks. The localisation of the sound design falls victim to being incorporated into a stereo mix. It’s sometimes evident to my ear that more movement in front of the right screen results in a faster, louder spark shower on the right hand side of the stereo mix, but I doubt many visitors will pick this out themselves. The ideal solution would of course be completely localised sound – one speaker (at least) for one screen. Still, for me the audio definitely works – it is essentially a development of the original idea I had of multiple spark noises in the foreground and a cycle of ongoing notes in the background to add some tension and sense of direction.

To improve the piece I would find a way to raise the spark threshold and ideally incorporate truly localised audio.


For me is a great success. Almost the last vignette to make it into the show and essentially the result of a happy accident. There is a fascinating moment when the piece begins that a few video frames are shown from the previous time it ran ie 10 minutes previously, resulting in the ghosts of people past being mixed in with those present. The particle + line effect sometimes doesn’t quite work as intended – somehow the lines move laterally rather than spiraling out in a more radial fashion – but this still looks good to me.  I’m also very happy with the audio side of the piece which is essentially a cacophony of staccato noise played over the top of a slightly plaintive de-synchronised generative  chord sequence. The effect is at once brash and yet almost dreamlike which somehow makes me think of the futility, and acceptance of that futility, of our every movement in this life.

A future improvement might be to deliberately store images on an infrequent basis and equally infrequently mix them into the randomly arranged sequence of recent video frames.

Sound x Light

Generally works as expected. Audio sensitivity can sometimes be an issue – either the piece is slightly too audio-sensitive or not audio-sensitive enough. This is probably the least exciting piece for me and feels to be somewhat of a ‘stocking filler’. But maybe that’s just because I’m too used to it now.

To improve it I would experiment with a range of mics and perhaps a limiter / compressor to enable sensitized audio response without over-sensitivity.

Data Flow

Needed significant tweaking to stand out from the screen – the original dark image with fine lines of light  just wasn’t coming through in the projected image. Still, I’m pretty happy with the result and would rather have the amended version in the show than not at all. The audio is murky in a good way as it helps to create an eery and otherworldly atmosphere but the positional definition is lost somewhat in the stereo mix, as with some of the other pieces.

To improve, I would perhaps try to find a way of keeping the finer lines and of course would prefer to use truly localised audio.

As a whole

I think that the work as a whole hangs together well and is a pleasant distraction from the outside world, as intended. Cycling the vignettes at 2 minutes duration each, making a total cycle duration of 10 minutes, seems to be about right.

The next three days look to be substantially busier so it’s good to know that the piece works and has already been working for some time.

Producing a decent quality video and photographic record is proving to be difficult but that is a different subject.

Leave a comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s