Q Estimation with Uncertain Frequency Information

In this paper, we present a new method for Q estimation in the presence of unreliable frequency ranges. Those unreliable frequencies can be caused by noise in distinct frequency ranges or destructive interference, for instance, from free-surface ghost reflections. We introduce the usage of noise-model or ghost-model-based weighting functions for spectral-division-based Q-estimation methods. The benefits of using noise or ghost models are demonstrated for a common-ray-parameter-based Q estimation on synthetic data. The proposed modifications are suitable for any frequency-based Q-estimation method.