In speech synthesis, a [[subjective test]] is a methodology to evaluate the performance of [[what is speech synthesis|speech synthesis]] systems that involves human listeners rating the quality and naturalness of the synthesized speech. These tests are typically conducted in a controlled environment, where listeners are presented with a set of speech samples and asked to rate them on a scale of naturalness or clarity. The two most common subjective tests to evaluate the quality of synthesized speech are: - [[mean opinion score (MOS)]], where the human listeners rate each sentence on a 1 to 5 scale. - [[AB test]], where human listeners choose which of two utterances they prefer. Subjective tests are important because they provide a way to assess how well TTS systems perform from a human perspective, rather than just relying on objective measures such as accuracy or error rates.