Mann stellt Plakat vor

The Quality of a Synthetic Voice

The way Internet users consume information has changed considerably. A couple of years ago there was still a “reading society”, but today it has turned into a “watching and listening society”. The triumph of video portals, such as YouTube, Vimeo and others, bears testimony to this trend. The company clipnow from Babelsberg has seized the opportunities this trend offered and developed a technology which uses text or image blocks for the fully automated creation of video clips with added sound.

Ralf von Grafenstein is a busy man: he is the founder, main shareholder and manager of the young company clipnow and therefore has a lot on his plate when it comes to managing its rapid growth. The team already consists of 12 members and it is currently moving into larger facilities in the heart of Potsdam’s district Babelsberg. “Every month or every two months a new co-worker joins the team because the number of our clients is growing by about 10% every month,” clipnow’s boss points out happily.

Everything started in February 2010 when the company was founded; back then it was called 45info. Ralf von Grafenstein and his associates had the idea of converting articles of the popular knowledge platform Wikipedia into video films of 45 seconds each. To this end they developed a technology that uses existing photographs, graphics and texts for the fully automated creation of moving images with added sound. Step by step, 800,000 videos were created in English and German.

The early stage fund and “R&D of SMEs” supported the business idea

Von Grafenstein took his “Wikipedia film adaptations” to various financing funds. In the end, and through Brandenburg Invest, the “BFB Frühphasenfonds Brandenburg” (“BFB Early Stage Fund Brandenburg”) was convinced of the potential and market orientation of the business idea and invested into the start-up. The company bmp AG was tasked with the fund administration. Thanks to the funding programme “Research and Development of SMEs”, 45info/clipnow was able to advance its semantic text analysis and to optimise the use of clips for desktop computers and mobile devices. Specifically the quality and intelligence of the synthetic voice needed to be adapted to the clients’ requirements. During a past project conducted for a large travel company the pronunciation of foreign proper names emerged as a problem for the automated speech creation.

One and a half years later they did it: von Grafenstein and his team created their own “voice”. At the same time they were able to link the video creation to the XML format, on which many websites are based.

Perfect for estate agents and car dealers

clipnow is successful at the market thanks to its “all-included service package”, which is very attractive especially to clients with ever changing products, such as estate agents and car dealers. These clients can change or complement their online offers at any time, and an updated video clip is being created in a fully automated manner almost in real time. Furthermore, clipnow creates a specific YouTube channel for every single of its clients, which serves as an additional marketing channel. “In comparison to its effect, the costs for this comprehensive service are extremely low,” says clipnow’s manager von Grafenstein.

60% of all the videos on the platform immobilien-scout24 were created by clipnow already. Numerous renowned brokerage firms now rely on the technology provided by the folks from Babelsberg. “Until the end of 2014 we would like to have 1,000 clients from the real estate sector. At the same time we are expanding our activities in the field of cars, and then we would like to go international,” von Grafenstein points out with confidence.

On its path to the future clipnow will still benefit from the funding opportunities provided by the state of Brandenburg. Ralf von Grafenstein knows that Brandenburg Invest’s consultants will always stand by his side.