Realistic, cutting-edge, emotional – these are just a few of the adjectives that operators of AI-based vocal generators use to advertise their services. At first glance, the offerings seem attractive: you can try many of them for free, have a variety of voices to choose from, and even with a paid subscription, you save money compared to using human artists. At first glance, everything speaks in favor of these services. In reality, however, the opposite is often the case…
Poor quality
In preparation for this post, I tried various online text-to-speech services, including Eleven Labs, Uberduck, Audimee, and Musicful. Trying to achieve a meaningful, usable result, I spent hours tweaking prompts and markups. I finally gave up: the voices often sounded synthetic, lacked flow and emotion, and the audio quality…
