Papercup, a startup based in the United Kingdom, has raised $20 million for an AI-powered dubbing program designed to help students learn about their expressive voices and expand into new languages. The funding will enable Papercup to broaden their offering in markets in which they know their technology works right.
Customers upload videos, choose a target language, and receive a synthetic version. Sky News, Discovery, and Business Insider have already used the technology. More than 300 million people saw Papercup''s videos in the last year.
The promise of AI dubbing
Markets and Markets forecasts that the global film dubbing market will grow from $3.1 billion in 2021 to $4.7 billion in 2028. Currently, the majority of dubbing is done through a manual and tedious process involving human translators and actors.
Papercup has developed a novel human-in-the-loop technology that allows it to achieve the same quality as dubbing studios at one-fifth of the cost. It hopes this will broaden the market potential in games, podcasts, corporate training, and audiobooks.
Deepdub, Respeecher, and Resemble AI are among the others who are interested in AI-powered dubbing. Apart from that, other firms like Synthesia are aiming to not only dub videos, but also synchronize lip movements in order to achieve greater realism.
Quality feedback loop
Verbit, a financial consulting firm, paid $250 million to expand its operations in a bid to expand its programming, which is already developing a better feedback loop that allows experts to deliver specific feedback to improve individual quality, but the training data.
The company has used a wide spectrum of third-party training statistics and has developed its own to produce a comprehensive catalog of training statistics spanning many ethnic backgrounds and languages.
Shemen said that developing workflows to achieve this at scale is a huge challenge.
Papercup has enlivened an in-house dubbing software stack to collaborate with partner transcription and service providers before the text-to-speech engine generates voices. This way, quality assurance teams can control this service to improve translation accuracy and add nuance. They also created a feedback loop from local listeners for continuously improving the quality and naturalness of voices.
The service uses existing AI engines for automated transcription and machine translation before Papercup employs its own text-to-speech approach to create new speech tracks. A human-in-the-loop approach allows a professional translator to perform a quality check, modify, and modify translation and speech to improve quality.
According to Shemen, this is the most effective way to use available AI technology, which we can build on to create a commercially viable, productized AI platform.
Papercups'' total has been raised to $30.5 million by Octopus Ventures, which has included Local Globe, Sands Capital, Sky, and Guardian Media Ventures, Entrepreneur First, BDMI, and a wide range of angel investors, including William Tunstall-Pedoe, the founder of Evi (now Amazons Alexa), and Zoubin Ghahramani, a senior research director at Google Brain, and a former chief scientist at Uber.
