A Content Adaptive Learnable Time-Frequency Representation For Audio Signal Processing https://arxiv.org/abs/2303.10446