Overview of this campaign

The voice is created with an AI-based technology trained on huge amounts of voice clips. We have cooperated with the text to speech-company Beyond Words, and colleagues in Aftenposten. These four principles have been guiding:

1) We want to provide users with a richer experience. That means, they should be able to choose how to consume the news and the content. This is a megatrend in the business, and we are now working on an automated video solution. The motto: We create the content, you choose how you want to explore it. A format buffet, if you like.    

2) We want to increase accessibility. People with reading difficulties will now have easier access to all our content. According to the organization Dyslexia Norway, 5–10 percent of the population has this diagnosis. 

3) We want to explore new media products. We are certain that cloned voices are only the beginning. Next will be an automated podcast and the possibility of stacking all of your content as playlists. We can call it spotification of news.

4) We want to connect with a younger audience. Man prefer listening over reading, especially at certain times of the day. 

A bit about the process of creating the voice of Bergen as well: 

The ideation: How to detect the perfect voice of Bergen? To decide, we engaged the newsroom in a huge audition. 15 reporters participated, and an internal jury decided. A bonus effect of this was a profound newsroom interest in the project 

The creation: To clone a voice has become fairly easy. Getting a cloned voice speaking in a distinct manner, and speaking as close to perfection as possible, is not easy at all. The owner of the voice, Eir Stegane, spent hours after hours in the studio reading voice samples. 4000 sentences, to be precise.

The iteration: This technology is rapidly evolving. The voice is not 100 percent perfect. We have hired a linguist to help, and also rely on feedback.

Results for this campaign

Is it possible to create a synthetic voice that perfectly captures the distinct accent of a city? This has been the ambitious goal of Bergens Tidende’s local audio cloning project. Now, every single article is enriched with this voice. Feedback from listeners suggests that they like what they are hearing.

It's all about identification and trustworthiness. Being a local paper, emphasizing the uniqueness of our surroundings is of great importance. Of course, our synthetic voice should reflect the way people in our city speak.

Perhaps the most valuable outcome of this project is our ability to quickly develop new media products that meet users' expectations, particularly younger audiences. We know that these groups increasingly prefer to choose how they consume news, and that the traditional article is under massive pressure. The media industry is engaged in a versioning arms race, and with its cloned voice, Bergens Tidende is positioning itself as a leader.

We use the cloned voice to accelerate product development. We are already experimenting with an automated afternoon podcast that delivers the best of our content to listeners. We have also launched Spotify-inspired playlists with CarPlay functionality, and we are currently working on automating the creation of short videos based on the cloned voice. And more innovations will follow as development speeds up. As a first mover, BT is well positioned for this brand-new audio future.

Our Voice of Bergen has introduced BT to a new group of potential users, and we have received excellent feedback from people with different reading challenges.

Are users actually listening? More and more—and we expect this trend to continue as we iterate and launch new products

Contact

To contact a company representative about this campaign, click here for the INMA Member Directory