Robustness of TEC speech watermarking to cropping and additive noise

Full text