Introduction to data-handling tools

In data sonification, while the input data can be thought of as eventually controlling the sound rendering, the transformations it has to undergo in the interim can be quite considerable. Such data processing can reasonably include multidimensional scaling, filtering and statistical analysis, which itself may become the subject of sonification.

Each input dataset can have potentially unique structural characteristics. Some common characteristics are multiple channels with various (and variable) amounts of noise and offset biases, massively paralleled, metadata-embedded and multiplexed arriving in real-time. Difficulties in displaying such datasets are compounded when they need to be buffered and streamed in non-real-time as is the case of multiple overlays of time sequences of different temporal compressions.

High-level tools for processing such data complexities are rarely, if ever, found in computer music environments, and even less likely if the input data is spatial rather than temporal. In fact, it was the need for sophisticated data-handling tools that didn't have to be written for specific composition tools, and then maintained across various hardware platforms and their operating system upgrades, that led to the idea of the SoniPy Project.

SoniPy divides tools for data-handling tasks into the following categories:

• data acquisition
• data analysis and generation
• data persistence
• data transfer