Selvas AI announced on the 3rd that its AI voice content creation service, ‘Odiro,’ has upgraded its ‘AI Audiobook Editor’ and ‘3 AI Voices’ by reflecting customer demands.
Selvas AI conducted this upgrade based on the results of a survey of 487 audiobook listeners at last year’s Seoul International Book Fair, actively incorporating their preferences and requests.
According to the survey, 149 respondents (31.2%) cited “being able to listen to more books as audiobooks” as the biggest advantage of AI voice audiobooks. This highlights listeners’ strong anticipation for diversification of audiobook content. Following this, 147 respondents (30.8%) valued “being able to listen to one book in various voices,” and 129 respondents (27%) preferred “fast provision of new audiobooks.” The most preferred audiobook genres among listeners were novels (37%, 166 respondents), history & culture (23%, 104 respondents), and humanities (18%, 79 respondents).
Odiro is equipped with AI technology tailored for audiobook production. It applies ‘AI document analysis technology’ and deep learning-based voice synthesis ‘Selvy deepTTS.’ The service consists of an AI audiobook editor that can generate AI voice audio sources and produce audiobooks easily and quickly in large quantities at about 10% of the cost of traditional audiobook production by simply uploading books.
In particular, Odiro’s AI audiobook editor’s ‘AI document analysis technology’ automatically creates scripts and applies AI voices when books in PDF, EPUB, or TXT formats are uploaded. It allows simultaneous adjustment of AI voice volume, reverb control, and proofreading for major, middle, and minor chapters of the book. Sound effects and background music can also be inserted. While traditional single-volume audiobook production takes about 320 hours, Odiro can complete production within 24 hours, making it over 90% faster. Without the need to worry about using complicated DAW (Digital Audio Workstation) programs, publishers, audiobook producers, and authors can easily convert books into audiobooks automatically.
‘Selvy deepTTS’ generates natural and fluent high-quality synthetic voices that capture human breathing, intonation, and emotions. AI voices are created by referencing narration characteristics of popular audiobook genres, enabling genre-specific customized narration. Among the newly introduced 3 AI voices, ‘Jaehun’ offers a calm narration style suitable for novels, while ‘Juhwan’ and ‘Changwook’ provide narration styles and voices appropriate for history and culture genres.
Kwon Hyukmin, Head of Odiro Business, stated, “Audiobooks are important content as they enhance the value of publishing content and provide a broad reading experience to readers familiar with audio content.” He added, “Odiro will do its best to resolve the publishing industry’s concerns about ‘high production costs’ and ‘workload’ of audiobooks to offer rich audiobook content to readers.”
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
