Fully Automated Sound Correction Based on Acoustic Environment Estimation

In general, television (TV) sound can be deteriorated according to the acoustic environment of the watching space. While some TV models provide a room correction function, it is reported that very few users are using this feature. This lack of use likely comes from an inconvenient manual procedure that should be conducted by the user. Therefore, we propose a fully automated sound correction method on the TV without any manual operation by the user. The proposed method enables the estimate of the acoustic environment by a machine learning algorithm. TV sounds are optimized using preset sound correction equalizers designed in advance. The evaluation results show that the proposed method effectively improves TV sound in various acoustic environments.