What does “content” mean in a world of text, graphics, audio, video, and even virtual reality – with many variations on each of those themes?
The question reminds me of a talk I gave years ago in which I was trying to explain the American magazine market to some Chinese businessmen. They thought it should be a simple matter to take a print magazine and turn it into a digital magazine. I said they were right, in the abstract, but in practice, not so much.
They were imagining a single content repository that fed both the print and the digital editions. But what usually happens is a complicated and tangled mess. The editor writes a story, then it goes to layout, then the editor makes some changes, then it gets cut further to make room for a photo, etc., so the version in the print magazine becomes very different from the version in this fictional content repository.
There are solutions to that problem – or half-way solutions, anyway – and I don’t endorse the idea of a single content repository going to both print and digital, but you get the picture, and the purpose of this story is to set that up that image in your mind for purposes of today’s topic, which is text to speech.
Some people would rather listen to an article than read it. But creating an audio version of an article requires a lot of production, and if an editor changes something, it has to go back through production again.
But what if there was a way to convert text to speech that bypassed all that, so that when an editor amended an article, the audio version was immediately updated?
That’s the point of an article Bo Sack distributed a few days ago, titled “Schibsted invests in text-to-speech to engage audiences.” They’re in Norway, and there’s been a huge increase in podcast listening over there, and Schibsted wanted to ride the wave. There’s a follow-on effect from podcasts that drives popularity of other audio formats. Even with subscription news services.
“You can read your newspaper,” Lena Pederson says, “but if you also can listen to the news, then it’s more opportunities in your life to use your subscription.”
That is, providing audio increases engagement.
Some people would argue that it would be better branding if the authors and editors read their own work. There’s a kind of authenticity to that, and some people think that sells. But not every writer has a good voice, and tracks recorded by humans require editing and such.
However, “Beyond Words,” which is the technology they chose for this venture, does allow them to clone a popular voice and use that in their text-to-speech conversion – which sounds pretty cool.
There’s also the issue I mentioned above about day 2 edits. Text-to-speech technology can make an audio version of the current version of the article, even if it gets updated.
I like what they’re doing here, and I particularly like two things. The audio is served from their website. That means they control access, and they have all the stats. When you put audio on a podcast service, you can’t limit access to subscribers, and you don’t get as much data on usage.
This is something to watch, or … jump into, if you’re so inclined.
Links
