Audio Spotforming Using Nonnegative Tensor Factorization with Attractor-Based Regularization [2407.08951]