The categories for this prototype were inspired by Russolo’s mention, in his letter to Balilla Pratella, of the poet F. T. Marinetti’s account of a battle in free words. Marinetti created this language for his poetry that would destroy syntax and allow imagination to flourish without constraints. Russolo marvelled at the onomatopoeic astuteness of Marinetti’s depiction, capturing the essence of the battle through its noise:
Fury breathless ears eyes nostrils open! load! fire! What a joy to hear to smell completely taratatata of the machine guns screaming a breathless under the string slaps traak-traak whips pic-pac-pum-tumb weirdness leaps 200 meters range Far far in back of the orchestra pools muddying huffing goaded oxen wagons pluff-pluff horse action flic flac zing zing shaaack laughing whinnies the tiiinkling jiiingling tramping 3 Bulgarian battalions marching croooc-craaac [slowly] Shumi Maritza or Karvavena ZANG-TUMB-TUUUMB toc-toc-toc-toc [fast] crooc-craaac [slowly] crys of officers slamming about like brass plates pan here paak there BUUUM ching chaak [very fast] cha-cha-cha-cha-chaak down there up there all around high up look out your head beautiful!
F.T Marinetti, quoted in Russolo (1986:26)
In distancing myself from semantic descriptors, I applied this idea of onomatopoeia to Russolo’s six families of noises (1986:28). Observing the different families, I conceptualised single sound effects, written only with consonants, that I felt encapsulated the family’s characteristic sound.
My intentions to utilise Russolo’s six families of noises came early in planning the project, although I have continually struggled with exactly how I could name categories without having a space on the Videodefunct platform to display the table detailing the different families. Referring back to the sound samples I had already composed for Flânear and deciding which families they fell into, I appropriated the noises within the samples to determine my onomatopoeic categories. This process of naming the categories came relatively naturally and I was content with the results.
The idea of a purely perceptual descriptor for the categories derived from my intentions to abstract the sounds from their sources, instead placing emphasis on the aural qualities themselves. Perhaps through nonsensical categories, users navigating Flânear would be encouraged to form their own connections and opinions of media content to the classification devices. The supposed absurdity echoes Marinetti’s free words; imagination without strings. The representation of a Melbourne CBD soundscape in this example transcends the physical limitations of strolling through the city. Instead, soundwaves and their formations become the paths to traverse; follow a rumble down to a screech and leave with a bang.
Looking at a screenshot from the Videodefunct category management page once fourteen discrete media artefacts (hereby referred to as posts) had been entered and categorised into the system instigates several reactions. Although satisfied that the categories were certainly appropriate to describe all of the samples, the sheer volume of posts in each category suggested that there was too much crossing-over for this taxonomy to coherently classify the posts.
Russolo himself said of his six families that it “encapsulated the most characteristic of the fundamental noises; the others are merely the associations and combinations of these” (1986:28). True as this may be, my use of the six families of noises was capable in demonstrating similarities between categories, but characteristic differences, which are just as necessarily required for a cohesive folksonomy, are notably missing. This trend was reiterated in the tags I chose for this prototype.
Fully aware of the manipulations I had put the audio samples through in ProTools, I hypothesised that indicating this through classification devices would allow users insight into how sounds were treated and thus what sonic characteristics represented the Melbourne CBD soundscape. I want it to be apparent that all the sounds were sourced directly from the city, despite not necessarily indicating the exact sound sources themselves.
The tags for this prototype were the various manipulations used in ProTools to create the different samples, here explained:
• High-pass filter: equalisation technique that mutes higher frequencies
• Low-pass filter: equalisation technique that mutes lower frequencies
• Reverb: echo effect
• Repetition: cutting and pasting sound events repetitively
• Layered: numerous tracks layered on top of each other
• Stretch: stretches a sonic event over a longer duration
• Shrink: shrinks a sonic event to a shorter duration
• Reverse: plays sonic event backwards
Although I found some validity in revealing to users exactly what processes the audio had been through, there were far too many crossovers in the finished prototype to indicate any reasonable differentiation between posts. This caused a problem by destroying any real sense of navigation as posts were too readily revealed. Visually, selecting clips on the videodefunct display often resulted in an almost identical selection of clips appearing.
This issue can unfortunately be attributed to two possible errors in my process: one being the choice of the tags themselves, but another also being the samples themselves and methods I took in creating them. Rather than seek retribution in redoing all the audio samples so as not to allow overzealous crossovers with this tagging technique, perhaps it would be better advised to move far away from tagging according to ProTools audio manipulation devices.
Although the classification devices utilised in this prototype does well to attempt to indicate the nature of Melbourne CBD sounds, it falls short of fulfilling a navigatable representation of a Melbourne CBD soundscape. However, where taxonomy failed, the visuals and titles flourished. My personal impression of Melbourne’s CBD is that of a young, fresh location with a distinctly quirky attitude. Titles for posts were chosen quite pragmatically, often referring to the visuals themselves (such as objects, text or activities captured) or source location details. What emerged, once testing the prototype, was an alternative narrative altogether separated from the audio content, created by the horizontal display of image and text. Random combinations of photos and titles began telling their own stories that complemented the audio and further reiterated my impression of Melbourne’s CBD and consequently how I was representing it.
Given that the categories and tags were audio-oriented, the text and therefore images did not stray, crucially to Flânear’s intentions, from representing a Melbourne CBD soundscape. Serendipitously, the written narration emulated the representational aesthetics of Marinetti’s free word poetry.
In conclusion:
- So long as at least either the titles, tags or categories were of an audio-nature (onomatopoeic or technical), the text in the Videodefunct display will still encapsulate the nature of soundscape, complementing rather than competing with the audio.
- It is imperative that I begin thinking of tags and categories as inherent in acknowledging both the similarities and differences between the discrete media objects (posts) archived. Focusing on this characteristic of folksonomy has the potential to assist the navigatable aspect of the archive in recreating a culturally relevant representation of a Melbourne CBD soundscape.
- The images already allocated to audio have not raised expected problems, such as whether the abstraction from the sound source makes the soundscape more nonsensical than representational. Whilst capturing ‘essences’ of Melbourne, the pictures tend to emphasise the predominance of the audio used rather than the images, which ultimately is my desired relationship between the two. Although on occasion the titles are influenced by the images, this has proved to be aesthetically beneficial to the overall effect of Flânear rather than distorting the audiovisual connections.
- Still images have enabled a greater quality of audio as opposed to using moving images. And as the above point indicates, emphasis is once again allocated to the audio rather than the image. However, at this point, my investigation remains concentrated on the classification of the audio data rather than the impact of the images on the perception and navigation of audio data.
