Detecting gravitational wave signals using a flexible model for the amplitude and frequency evolution

We currently lack good waveform models for many gravitational wave sources. Examples where models are lacking include neutron star post-merger signals, core collapse supernovae, and signals of unknown origin. Wavelet based techniques have proven effective at detecting and characterizing these signals. Here we introduce a new method that uses collections of evolving amplitude-frequency tracks, or ``voices,'' to model generic gravitational wave signals. The analysis is implemented using trans-dimensional Bayesian inference, building on the earlier wavelet-based BayesWave algorithm. The new algorithm, BayesWaveVoices, outperforms the original for long duration signals.