As global content consumption increases and demand for non-English content exceeds that for English movies and series, IMAX is using AI to scale the localization of its original content.
The entertainment and media industry grew 5% to $2.8 trillion in 2023, in line with a report from PwC. The industry is anticipated to proceed its expansion, albeit at a modest compound annual growth rate of nearly 4% to $3.4 trillion over the following five years. Non-English language content is can also be growing rapidly on the English marketsincluding the USA, Great Britain, Australia and Canada. Last yr, Netflix reported a 90% growth in viewership of non-English content within the UK within the last three years.
IMAX has taken all of this into consideration and is now exploring localization using AI to draw more attention.
On Monday, the Canadian production theater company known for its massive cinema halls and immersive film experiences announced its partnership with a Dubai-based startup Camb.ai to make use of its AI language models to translate original content, including documentation.
Camb.ai, which has already used its AI dubbing and language translation for live sporting events and leagues, including Australian Open, Eurovision SportsAnd Major League Socceroffers its Boli model for speech-to-text translation and Mars for speech emulation. The models can be found through the startup's DubStudio platform, which supports 140 languages, including various low-resource languages ​​for which no significant data is accessible on the Internet.
“Companies like OpenAI and Anthropic have a unique vision of society,” said Akshat Prakash, co-founder and CTO of Camb.ai, in an exclusive interview. “They attempt to create models which can be very horizontal and might cover a big selection of tasks. …We don’t have to try this in any respect. Some of our models contain fewer than 100 million parameters and are highly specialized.”
Prakash, a former Apple engineer who worked with the team that developed AI and ML models for Siri, co-founded Camb.ai together with his father Avneesh Prakash last yr.
“We are a generation apart and grew up in India. “30 years apart, we faced the identical language challenges,” the CTO told TechCrunch.
He said Camb.ai pre-trained 70% of its models on academically licensed datasets which can be commercially viable. The remaining 30% involves fine-tuning the information the corporate receives from early partners using its models for AI-based synchronization and translation.
“What we are usually not doing, and we have now been very careful and have completely avoided doing this, is possibly abolishing the web,” Prakash claimed. “Some corporations think they will get away with it because they're constructing a consumer-facing app or tool and think it's OK to scrape about 10 petabytes of the web.”
Camb.ai uses a “three-tiered” approach to providing AI-based translation, which incorporates the bottom layer of its Boli and Mars models, the infrastructure layer that hosts these AI models, after which the DubStudio platform for the front end.
Unlike other AI-based models, Camb.ai's Boli takes input speech tokens, produces output text tokens within the translated language and preserves nuances, claims Prakash. Once Boli generates the text, Mars translates the text into speech using the identical audio input signal to capture the performance of the particular audio, including ambient noise, comparable to the background music of cheering crowds at sporting events, he said.
Prakash told TechCrunch that Camb.ai's technology delivers voice translation in as much as 10 languages ​​concurrently with a latency of 20 to 30 seconds, which could be covered by the streaming and broadcast delay of 30 to 40 seconds.
IMAX will introduce AI translation regularly, starting with resource-intensive languages. The deployment occurs after internal testing of Camb.ai's technology against its original content.
“Although we’re only within the early stages of the partnership, we are going to proceed to work together to higher explore its potential and the way it could actually best move us forward,” said Mark Welton, President of IMAX Global.
Welton said using AI would help save translation costs, without disclosing details.
Camb.ai currently has a team of fifty employees. In February, the corporate raised $4 million in a seed round led by Courtside Ventures. Prakash told TechCrunch that the startup is closing a bigger pre-Series A round to expand its reach and headcount.
