Q] You are using generative AI to create seamless dubs which help in keeping it real and also save costs of reshooting the ad for a different market. Could you tell us about its origins and the services that you provide for the advertising and film industry?
The starting point comes from our intense love for watching content, irrespective of the language that it is created in. Be it ‘Money Heist’ or ‘Squid Games’, we have had wonderful shows that got dubbed into the language of our choice. However, watching dubbed content eats away the sense of realism and authentic cinematic expression, as the audio does not match the visuals. Which is why we decided to deploy generative AI to fix this visual discord. The obvious use cases are in films and shows across various platforms. But this tech has a lot to offer to the advertising industry as well. In India, over 95% of the ads are shot in Hindi, and then they are dubbed into multiple languages. For regional customers, this communication is alien; it looks fake and unauthentic. As brands, you aim to authentically connect with customers and yet, at the same time, your communication is not authentic. To typically solve this using existing resources, means creating content in every language, which swells the production cost. Brands have lived with this problem for a long time, but now with the help of VisualDub, they can shoot an ad film in one language and then authentically convert it into the language preferred by the consumer. Whether it is films or streaming platforms or advertising, VisualDub is the perfect technology to help you achieve it.
Q] Please tell us the reason behind naming your startup Neural Garage?
Neural Garage is a composite of two words. First being ‘Neural’ because of the nature of our work. Neural networks are at the heart of all technology built around AI. The second one is ‘Garage’ because we wanted to pay tribute to big tech startups such as Microsoft, Apple, Google and Facebook that emerged from garages. We aspire to be placed among the biggest tech companies in the world. Hence, we felt that there could be no better name than Neural Garage.
Q] The leadership at Neural Garage seems diverse, with you rising through the ranks from the media and entertainment industry, and your co-founders hailing from the IT industry. How was this transition process for you?
Working hard year-after-year for other companies creates this urge to create your own legacy. But it has to be something that is compelling enough for you to leave behind your steady corporate career and get into this whole unknown adventure of entrepreneurship. The leap can be made if the project is compelling enough, you hold some passion towards it, and most importantly, you should believe that you can contribute and build a legacy around it. My co-founders are very bright engineers from IIT Kanpur, with a background in facial recognition. They have already built a company and sold it to a large US-based firm. I met them through some common network when they had already started working on VisualDub, the tech that we use. As the founding team, we felt that we could bring strong expertise in both technology and the domain for which we are building it. The goal was to build a business that creates noteworthy technology and services for three large industries.
Q] We were all in awe of Cadbury’s hyper-personalised ‘Shah Rukh Khan-My-Ad’ campaign that won big at Cannes Lions last year. Your project is similar to that but it adds multiple more layers of technology. Could you go into the workings of it?
In the Cadbury ad, just one word from the entire scene was replaced, which was done using text-to-lips technology. In text-to-lips technology, word changes happen through a text input, which gets converted into speech, and then it gets lip-synced. What it can’t change though is facial structure to match the new language. Our technology uses audio as a controller to change the spoken visual expression, which is a core part of our patent. When we speak, multiple facial features move: lips, mouth, chin, neck, jaw, smile lines, and muscles around the eyes. Our technology transfers all this information to match a new language, which is a complex task.
Q] Not so long ago, Neural Garage came into the light for your project with Amazon featuring Manoj Bajpayee. Could you name some of your other big clients?
Some of our top clients include Microsoft, Amazon, Coca Cola, Britannia, Ultratech Cement, and Dream 11. On top of that, there are certain other big names like GSK, HP, and 99acres.com that are testing our technology. I am extremely delighted by the fact that in such a short span of time, we have received validation from some of the biggest brands in the world. And this is a testament to the fact that our technology delivers to the expectations of our consumers.
Q] Could you mention some of your collaborations in the space of cinema?
Very recently, we signed a partnership agreement with UFO movies. UFO movies is India’s largest theatrical distribution partner. Effectively, every film released now goes through their pipeline, giving us a great channel partner and ensuring our technology is distributed. This brings our technology closer to consumers. Several large studios are testing our technology, but I can’t name them due to NDAs. However, if you look at the top film producers and major films being shot right now, you will find that we are being tested by them. That is all I can say for now.
Q] From my understanding, I don’t think you have competitors in India as of now. But what is the situation internationally?
In India, there is no competition, but on a global scale, we have two competitors. One is US-based Flawless AI and the other is Marz, which is Canada-based. They are solving the same problem, but with the key difference being that their heavy lifting is done by VFX, which is followed by a layer of AI, while ours is done primarily by AI. Additionally, they use third-party AI software or research, whereas we have developed everything in-house, from research to algorithms, making it all Neural Garage IP. This is significant because while India excels in products and services, it’s rare for an Indian deep tech startup to build software entirely from its own research. VisualDub is a truly made-in-Bharat technology poised to serve global consumers dominantly.
Q] What lies ahead for Neural Garage in terms of the vision, and will we be also seeing some major upgrades?
The vision for Neural garage broadly is to make communication seamless. Language should not be a barrier to understanding anyone absolutely authentically. In terms of features, we will be sharing this as market knowledge soon, but we have also developed voice cloning technology in house that becomes a great addition to our consumers. Imagine a Shahrukh Khan film dubbed in Tamil with a voiceover artist that doesn’t sound like him. With our voice cloning and lip sync technology, we can solve this. After dubbing, we can adjust the film so it looks and sounds natural in the new language.
