An Echo of What Once Was

A game made by Caleb Zarko.

The original concept of this project was to experiment with synthetic generative audio via Pure Data and integrate it into Unity. More specifically, the audio developed was done so to provide ambient noise to a scene in two different manners. The first category of sound would be ambient noises that have adjustable settings to allow for modification of the audio in real time to enhance changes to environmental settings in-game. For example, rain and wind sound different inside a house versus outside a house. The second category of sound would be that of spatial ambient noises, in this case that of fire. Unlike rain and wind, which are pretty omnipresent in an environment, fire is something that can be approached or receded from and thus this synthetic audio has to be able to detect player proximity by some means. The current state of the project is not perfect by any means, as this is new content being learned, but is at least a presentation of potential.

In reflection, due to time constraints in addition to underestimating the complexities of Pure Data, one of the key changes, from the initial proposal of this project at the time, is a reduction of the total produced sounds. In the current state, there are 3 sounds produced using Pure Data, wind, rain, and fire, but originally the intention was to produce roughly 5 sounds, including that of thunder. It can also be argued, that the original proposal had a focus on the fidelity of the synthesized audio as well, but, again, underestimating Pure Data meant sacrifices were made in order to produce a few sounds rather than a single higher quality sound. Despite these difficulties, however, I was able to develop a better understanding of Pure Data and familiarize myself better with the various short cuts and tools of the vanilla program. Overall, the obvious benefit of this project is that now, knowing Pure Data better, I can expand on this project or commit to other projects and produce greater results. While a simplistic takeaway, in some senses, it is the most important on the path to developing the skill set to create generative ambient noise that is believable. In a deeper sense, one of the greatest takeaways from this project is growing understanding of how mathematical sound is. In the process of learning and understanding Pure Data, I found myself occasionally reading scholarly articles that deliberated on the deconstruction of audio into component parts and the formulaic representation of those components mathematically. In regard to peer feedback on this project, input was somewhat limited. Developing this in a Music/Sound oriented class, meant that perhaps pursuing a project such as this meant exploring a skillset and field relatively untouched among my peers. One feedback that was received, for example, regarded the sound of the rain which sounding synthetic can make it a little hard to differentiate from the other sounds at times. While true, and something I would love to fix, it's not something possible to improve significantly within the allotted time. It would require a little bit of a better understanding of Pure Data in addition to the mechanics of real life rain noises. That being said, their encouragement and support during the process was reassuring. In the future, if I were to approach a project similar to this again, I would focus my scope on producing a single synthetic sound and substitute additional environmental sound with sampled audio. The reason for this being that the sampled audio, being easier to modify and produce environmental sounds for a game, would allow me more time to focus on the quality of the single synthetic sound I would be producing. It can not be overstated that the decision to make multiple sounds in the proposal became an unfortunate restriction in some sense of the word. In the end, this project was a fantastic means of exploring this concept, and I am very excited for future projects that will employ Pure Data.

Controls

• 'A'+'D' + MOUSE LOOK - movement.
• SPACE - jump.
• LEFT CLICK - shoot.