In a nutshell: Spotify studies that the variety of individuals listening to podcasts has surpassed 100 million and continues to develop. The streaming service is now exploring new strategies to broaden entry to its podcast service for a fair bigger viewers of creators and languages. It goals to realize this by leveraging the capabilities of AI algorithms to robotically translate content material from some outstanding figures within the business.
Spotify has unveiled Voice Translation for podcasts, a brand new pilot service that makes use of AI to translate audio reveals into varied languages. Via the applying of machine studying strategies, podcasts can preserve the unique podcaster’s voice, leading to a extra pure sound.
Voice Translation is constructed upon a instrument developed by Spotify and incorporates cutting-edge developments in AI audio expertise, together with OpenAI’s contributions. OpenAI, headquartered in San Francisco, has developed a flexible speech recognition mannequin referred to as “Whisper,” which has undergone coaching on a considerable dataset of numerous audio. This coaching allows it to carry out duties equivalent to multilingual speech recognition, speech translation, and language identification.
Because of Whisper and its AI mannequin, Spotify claims that translated podcasts sound extra pure than typical dubbing. In response to the streaming service, a podcast recorded in English can now be provided in different languages whereas preserving the speaker’s “distinctive speech traits.”
The preliminary assortment of AI-translated podcasts, compiled on a devoted Voice Translations Hub, options episodes from well-known podcasters equivalent to Dax Shepard, Monica Padman, Lex Fridman, Invoice Simmons, and Steven Bartlett. At this stage, solely a restricted variety of catalog episodes and upcoming releases will bear translation. Spotify has dedicated to offering translations in Spanish, French, German, and different unspecified languages. The corporate can be collaborating with different podcasters to incorporate their reveals within the translation service.
In response to Ziad Sultan, Spotify’s VP of personalization, providing translated podcasts that retain the creator’s authentic voice allows listeners worldwide to find and be impressed by new podcasters in a extra genuine method. Spotify is taking a considerate method to AI to foster a deeper connection between listeners and creators, which seems to be a key factor of the corporate’s mission to unlock the potential of human creativity.
Voice-translated episodes shall be accessible worldwide for each premium and free subscribers. Initially, translations shall be obtainable in Spanish, with French and German variations set to be launched within the coming days and weeks.