The voice effect or principle assumes that people learn better when they are exposed to multimedia instruction that includes a human voice rather than a machine voice. This investigation reconsiders the voice principle by investigating the relationship between learning outcomes and mental effort of the learners. Text-to-speech (TTS) engines have improved dramatically since the early 2000s, thanks to technological advancements. The researchers employed sophisticated TTS engines in a pretest-posttest design to analyze the various voice types (human voice, traditional machine voice, and modern machine voice). The results indicated that the progress in TTS technology enabled to generate real-life-like voices, and therefore no significant difference was observed between the modern and human voice. Furthermore, the participants' cognitive load was consistent with the findings of the learning outcomes.