VC-PWQ: Vibrotactile Signal Compression based on Perceptual Wavelet Quantization

The field of vibrotactile signal compression has gained significant traction recently. The aim of vibrotactile codecs is to transmit vibrotactile stimuli across distance with minimum delay and bandwidth requirements, while maintaining high signal quality. Several vibrotactile codecs have been proposed in the past that have shown the potential of such methods. In this work, we introduce an improved vibrotactile codec that builds upon one of the state-of-the-art codecs. By enhancing and optimizing core components of the codec, we are able to substantially increase its performance. In particular, with the employment of a more accurate perceptual model, better quantization method and arithmetic coding, the proposed codec significantly outperforms the state-of-the-art with respect to both objective and subjective quality metrics. Thus, we are now able to compress signals even further, while maintaining perceptual quality for the human user. This is particularly important for multi-point interaction scenarios where the raw data rate quickly grows with the number of interaction points. The implementation of the proposed codec is to be made publicly available.