raiNdrOps mISalignEd is a live performance that deconstructs a field recording from the semi-tropical environment of the Hong Kong archipelago, specifically the dense sound of forests on Lamma Island including a noisy mix of insects, bird calls, and rain. The deconstruction reconstructs its content using sine wave oscillators. The resulting remediations are spatially dispersed while the original recording is heard in stereo as it was first captured.

Through sound conventionally associated with the domain of nature, the composition engages with the synthetic listening practice of a computer and thematizes resynthesis, which depends on the synchronization between time and spectral domain information. After overcoming all technical barriers, resynthesized sounds transcode as the exact copy of the original recording. Machine listening can be described as interference with human listening because it obscures the process of remediation.

raiNdrOps mISalignEd, 2025, digital audio & resynthesis with projection; composition for eight-channel system; 12 minutes.

raiNdrOps mISalignEd, 2025.
The performance alternates between the original field recording from Lamma Island, Hong Kong, and its resynthesized counterpart, producing moments of perceptual interruption that expose the technical mediation of listening and draw attention to spectral representation as an operative rather than merely descriptive process. The audio is analyzed in a custom-built Max/MSP patch, where a fast Fourier transform (FFT) converts the signal into discrete frequency bins. The resulting complex spectral data are transformed from Cartesian to polar coordinates to extract the magnitude and phase of each spectral component, then routed to an oscillator bank. Oscillator frequencies are determined by the FFT bin index and the frequency resolution defined by the FFT size and sampling rate. As the oscillator bank reconstructs only an approximation of the analyzed signal, the resynthesized audio differs subtly from the original despite preserving its overall spectral character.