Sonification Design

The following paragraphs will describe, how the sonification and sound design for the first prototype could be approached at this point of the research process:

The sound design of the first prototype will make us of a mix of earcons and continuous sounds created through parameter mapping. Artistically speaking, the planned soundscape will sound like an artificial motor or electronic heart, that represents the flow of the business. Whereas "positive metrics" (such as data successfully uploaded) will preserve the harmony and rhythm of the soundscape, errors and failures in the process and system will create sounds that disturb that harmony.

Below is a list of all metrics that will be sonified in the first prototype and what they represent for the business:

Query respond times

DataShaka continuously runs queries to all databases used by their clients and measures the response times to constantly see the performance of this account. Through that process, employers are able to view query respond times, review and compare historical query response data and most importantly see if the database is completely down. Those speed queries will be sonified with earcons. The planned sound design will be similar to a sonar, possibly adding a constant rising sound until the response comes back. This way, exceptionally long responses can be spotted inside the continuous soundscape, as the noise until the response increases.

Moving UPDs

UDP (unified data point) is a unit to measure the amount of data. It represents one single identifiable point of data in the file. This way of measuring amounts of data is more accurate than referring to the file size, as this unit is file format independent. 

UDPs are moving through different steps of the process and go through various checks before being uploaded into the database system. Failures in those steps will be represented by recognizable earcons disturbing the harmony of the overall soundscape.

When UDPs are being uploaded into the data platform successfully, the process will acoustically represented the process with constant sine waves creating a so called Shepard Tone[1]. This tone generates the auditory illusion that its frequency is continuously ascending (or descending) though it actually remains in the same frequency range. The duration of this sound is dependent on the amount of UDPs being uploaded.

PL1 Health

A cloud machine (pl1) is used by DataShaka to process most of the data. It's "vital signs", such as CPU, Memory Usage, Network Usage, Disk Usage and Space etc. are highly important, as the data flow of the company depends on it. To be aware how busy and healthy the cloud machine is, all those values will be connected to constant wave generators. The waves will become louder and more intense as the values increase. The sounds will be designed in a way, that each of these signals will be identifiable, so it is easy to figure out which parameters might become critical.

Number of Jobs

The amount of jobs being processed on the cloud computer is a relevant number for the business as well. It is important to be able to distinguish between jobs awaiting to be processed in the backlog, jobs having finished the process, jobs getting stuck during the process and jobs being processed as we speak. The amount of jobs being stuck as represented similar to the pl1 health with an increasing wave that becomes more recognizable the more pressing the problem gets (in addition to the earcons that are triggered when the job gets stuck). Similar approaches will be done for the amount of jobs being processed (in addition to the earcons that are triggered during particular steps in the process). Jobs being pulled from the backlog or pushed into done will be sonified through auditory icons or earcons.

In summary, Speed query responses and Moving UDPs will be sonified using earcons/auditory icons whereas the cloud machine metrics and the number of jobs will use parameter mapping. 

As all data is time series data, Model-Based Sonification will probably not take a major role in the sonification approach. However, since the prototype will definitely provide interactivity, e.g. to enable users to focus on particular metrics, ways on how to include this exciting sonification technique will be explored.

All these sound design approaches will be tested and evaluated for their specific use cases. This will be a major part during the first user studies, where this prototype will be put to action. 

References

Shepard, R. 1964. Circularity in judgments of relative pitch. The Journal of the Acoustical Society of America, 36 p. 2346.