Humelo, a South Korea-based startup that specialises in AI-based voice synthesis solutions, raised $2.6M in pre-series A funding.
The startup’s AI TTS (text-to-speech) solution, Prosody, can create AI voice-read sentences from a short sample audio recording. Prosody’s most defining trait is its ability to create human-like voices that have variations in intonation (through adjustment of tone, speed, etc.) and convey emotions that are appropriate for the situation at hand—an effect that is the equivalent of using a voice actor. Furthermore, compared to existing technologies, which require a sample recording of at least 30 to 60 minutes long in order to combine voices, Humelo needs a sample that can be as short as only two minutes.
Humelo’s technology allows users to convert a Korean language recording into other languages, including English, Chinese and Japanese. The users can also convert English, Chinese and Japanese into other languages as well. Users can also isolate one voice from an audiovisual file and change it into a completely different voice of one’s choosing.
The startup is currently engaged in diverse voice synthesis collaborations with notable Korean companies, such as KT and SM Entertainment, based on the growing recognition of its AI voice synthesis technologies.
Humelo CEO Lee Ja-ryong said:
“The market for AI-based voice synthesis technology, which is making forays into the fields of entertainment, personal broadcasts, audio books, and IoT devices, is essentially limitless. Going forward, we will be concentrating on developing new technologies and services that will allow Humelo to secure a dominant position, technologically, in all of these areas and more.”
Humelo will be using the funds for aggressive hiring and the development of new services, such as dubbing in multiple languages and creating voices based on voice synthesis.
