Deepdub’s Oz Krakowski Talks Expressive AI Dubbing, Global Monetization, and Live Translation at Scale
In this interview with Streaming Media contributing editor Jan Ozer, Oz Krakowski, Chief Business Development Officer at Deepdub, discusses how the company delivers emotionally rich AI dubbing for media and entertainment clients. From crime documentaries to sports and theatrical releases, Deepdub’s hybrid approach blends machine learning with creative talent to localize content at scale while maintaining quality and nuance.
Krakowski outlines four engagement models—ranging from fully managed service to API-based access to a newly launched live translation platform. He shares how studios are monetizing long-dormant libraries with AI dubbing, how region-specific audience expectations shape strategy, and why Deepdub’s expressive voice capabilities and security credentials have made it a go-to partner for Hollywood studios.
Below is a lightly edited version of the conversation.
Jan Ozer: I'm here with Oz from Deepdub, to discuss Deepdub’s namesake dubbing capabilities. Thanks for joining me, Oz.
Oz Krakowski: Absolutely. Thanks again for having me.
What Deepdub Does
Jan Ozer: What does your company do? What's the value proposition you bring to your customers?
Oz Krakowski: Deepdub is heavily invested in the media and entertainment business. Our core is AI voices that have full expressivity and emotive expressions. We take audiovisual content from one language to another—and a few other things we’ll talk about.
Legal vs. Audience-Driven Localization
Jan Ozer: Walk us through the localization landscape. Some of it is caption only, some is audio. How much is legally required, and how much is driven by monetization?
Oz Krakowski: It's very region-specific. More than legal requirements, it's about audience requirements. For example, in Latin America, if you're only offering captions, your audience is going to be 1%. It's very minimal. The only way to reach audiences is through dubbing into local languages. Otherwise, you're just not going to get the traction you need.
Clients and Monetization Strategy
Jan Ozer: Tell us some trophy customers you can talk about.
Oz Krakowski: We’re speaking onstage at NAB today with Paramount. They’re using our platform and tools to do some of the work themselves. Other customers include Forensic and FilmRise—we dub docu-crime content for them. We’re working with the majority of Hollywood studios, and we’re the only ones who can say that. We help them monetize content—and also help major independent studios monetize content that would otherwise sit in a basement doing nothing.
Jan Ozer: What are the economics of that? What does it cost to convert an audio track and how much extra money do they make?
Oz Krakowski: It’s very dynamic—region-specific, language-specific, genre-specific, and even requirement-specific. A theatrical release will be different from something that goes to YouTube or a FAST channel. The beauty of what we do is we know how to balance it.
We’re not just a tech company. About half our staff comes from the creative industry—directors, voice actors, sound engineers. We know how far technology can take us and where to bring in creative professionals to meet the customer’s quality threshold.
Differentiating from Other Voice Platforms
Jan Ozer: It seems like there are a lot of companies offering expressive voiceover. How do you differentiate?
Oz Krakowski: Multiple fronts. First, we were among the first in the space—we started in 2019—which gives us a lot of experience. The fact that we’re working with major studios is a testament to that. We check all the boxes: security, legal, ethical, voice acquisition. Studios know they’re getting a high level of assurance when they work with us.
Beyond that, we designed our voices specifically for this industry. They have full expressivity—they can whisper, shout, scream, giggle, be sarcastic. And we have the infrastructure and people to make sure what we deliver meets their quality bar.
Four Modes of Engagement
Jan Ozer: How do you make your tools available to customers?
Oz Krakowski: Today we offer four major ways to engage.
- Fully managed service: Give us your content—we handle everything. Whether it's a single season or 5,000 episodes, we can scale.
- Licensable platform: Our GO platform is fully licensable. We'll train you, and you can use it in-house for presales, marketing, or any part of your pipeline.
- APIs: If you just need to tap into our engine—for example, to localize audio descriptions into other languages—we can support that via API. A good use case is accessibility compliance with the European Accessibility Act in June.
- Live dubbing: We just launched this last week. We take a live stream—HLS or SRT—transcribe, translate, generate the target language, and send it back to the broadcaster.
Jan Ozer: Is that with full expression?
Oz Krakowski: It’s with full expression. It’s especially fun to watch with sports broadcasts—you get all the energy of the commentator.
Jan Ozer: That works for sports, where the speaker isn’t on camera. What about talk shows or when the person is visible?
Oz Krakowski: Right now, we aim for what we call “UN-style.” It’s voiceover. You duck the original and overlay the translated voice. We don’t change the visuals—so no lip-sync—and there’s a short delay. But in markets like Latin America, France, and Germany, that’s a totally acceptable format.
Latency and Future Goals
Jan Ozer: What’s the latency for a live event?
Oz Krakowski: Right now we’re at about 20 seconds. That allows us to maintain context, even for long sentences like this one. The ability to bring that down to 10 seconds is there, and we expect that to happen in the next couple of months.
Related Articles
With key streaming services such as Disney+, Amazon, and Netflix trying to drive down production costs across the board, premium content providers have spent considerable time looking at how they can develop or license content which isn't produced in English but can offer global appeal.
26 Aug 2024
Today, localisation remains a critical budgetary line item for content owners delivering shows to diverse and transnational audiences, and it is probably one whose typical costs have not, until recently, changed considerably in quite some time. The increasingly prevalent use of AI in content localisation, subtitling, and translation promises to change all of that—particularly through the controversial and ethically fraught use of imitative synthetic voices.
26 Aug 2024