Skip to content

Blog

From re:Invent Spotlight to Global Impact: MagellanTV's Journey with Mission

From re:Invent Spotlight to Global Impact: MagellanTV's Journey with Mission | Mission
6:04

 

At AWS re:Invent 2023, MagellanTV's innovative AI-powered dubbing solution took center stage during Dr. Ruba Borno's Partner Keynote, showcasing how collaboration between Mission and MagellanTV is revolutionizing global content delivery. 

As we look back at this milestone achievement and examine MagellanTV's progress since then, it's clear that this was just the beginning of a transformative journey in documentary content distribution.

The Road to re:Invent 2023

What began as a conversation about cloud cost optimization became much more ambitious.

When MagellanTV initially approached Mission, they sought expertise in managing their growing AWS infrastructure. As Mission's team dove deeper into understanding MagellanTV's vision, they discovered an opportunity to revolutionize how the streaming platform delivered content globally with generative AI.

MagellanTV faced a significant challenge: how to make their vast library of over 3,000 documentary series accessible to audiences worldwide. The traditional approach to dubbing and translation was both time-consuming and costly, with prices ranging from $20-25 per minute of content. The manual editing process could take up to 60 hours to dub just one hour of content, consuming approximately 70% of their video editors' time.

Mission's solution leveraged AWS's powerful suite of AI services to create an automated pipeline dramatically reducing cost and time investment. 

By combining Amazon Bedrock, SageMaker, Transcribe, Translate, Comprehend, and Polly, the team developed a sophisticated system that could handle the complexities of documentary content - from managing background sounds to dealing with overlapping voices and maintaining proper timing.

The results were game-changing for MagellanTV. 

Dubbing costs plummeted to under $1 per minute, and the automated pipeline significantly reduced the manual effort required. This transformation caught AWS's attention, leading to the solution being featured at re:Invent 2023's Partner Keynote, where Thomas Lucas, Co-Founder of MagellanTV, shared the stage with Simon Anderon, Mission CEO.

Watch the keynote here!

The Sequel: MagellanTV's Post-re:Invent Production

Since their re:Invent appearance, MagellanTV has been actively expanding and refining their AI dubbing capabilities, with Mission's team providing continuous engineering support and expertise through their managed data ops services.

At AWS re:Invent 2024, we had the chance to catch up with Andrew Federowicz, Director of IT and Product Lead for Magellan Voice Works at MagellanTV. Andrew shared with us that the company, in collaboration with Mission, has been focusing on several key areas:

Pipeline Robustness 

"Over this last year, our focus has been fleshing out the pipeline and making sure it's robust," Federowicz explains. Mission's engineers have been working alongside the MagellanTV team to ensure the system can handle increasingly complex content, including documentaries with challenging audio conditions. 

Language Model Evolution 

The rapid advancement of AI technology has played a crucial role in their progress. "We started with earlier models a year ago. Now we're at Claude 3.5," notes Federowicz. "That in and of itself has completely changed the landscape of how these AI models translate and process words and shorten sentences." Mission's team has been instrumental in helping MagellanTV adapt to these evolving technologies, ensuring they stay at the cutting edge of AI capabilities.

Expanding Language Capabilities

For MagellanTV, reaching English-speaking audiences was just the first chapter. “Our goal is to be able to offer content in ideally dozens of languages,” explains Federowicz. 
“With approximately one billion English speakers worldwide and over seven billion people who don't speak English, the potential for growth is immense. Each new language represents an opportunity to share compelling documentaries with audiences who previously couldn't access this content."

Now Streaming: MagellanTV’s Global Future

The success of the AI dubbing solution, Magellan VoiceWorks, has opened doors for MagellanTV. 

Content providers have taken notice and are reaching out to explore possibilities for translating their content into various languages. Even challenging scenarios, such as documentaries with complex audio environments, have become opportunities for innovation and improvement.

MagellanTV's journey since re:Invent 2023 exemplifies how technological innovation can break down language barriers and democratize access to educational content. With their continually evolving AI dubbing solution, they're making knowledge and stories accessible to audiences worldwide.

Throughout this expansion, Mission's managed data ops team has been instrumental in accelerating MagellanTV's progress. "Working with Mission and Mission's engineers and project managers... it's like you have two arms and then you suddenly grow six more," explains Federowicz. "You have this ability to start developing stuff with people who are absolutely brilliant in fields that normally you'd have to spin up an entire division if you were working within your own company."

The collaboration between Mission and MagellanTV demonstrates the transformative power of combining AWS technology with innovative thinking. As MagellanTV continues to expand its language capabilities and refine its processes, they're setting new standards for what's possible in global content distribution.

For MagellanTV, what began as a simple cost optimization partnership evolved into an innovative technical solution showcased at re:Invent and a cornerstone of their global expansion strategy, promising to bring compelling documentary content to audiences in every corner of the world.

 

Author Spotlight:

Mission

Keep Up To Date With AWS News

Stay up to date with the latest AWS services, latest architecture, cloud-native solutions and more.