resize text:

Online Journalism News

Text-to-speech podcasts are a step backwards

Posted: 12/04/06 By: Jemima Kiss
email this story | post a comment

This is intriguing; a US local news site has an automated service to convert text articles into podcasts.

Sounds great? The Arizona Republic site has just won a Presstime Magazine award for the innovation, and I can see why this makes sense. It is an efficient way of re-formatting content for a different platform. All the text content is there, so why not develop an automated way of repurposing it? Everyone's happy!

However...

Anyone that has ever fiddled with the 'start speaking text' tool on their Mac or similar tool on their PC will know what a strain it is to listen to automated text. It is incredibly hard on the ear because it lacks the natural, sympathetic undulation of real speech. Your ears get the same kind of unwelcome work-out they would get with an automated phone system: "Extension [pause] one [pause] five [pause] nine [pause] six [pause] is unavailable!"

In addition, there is a whole craft of broadcast journalism skills that really need to be applied to audio content. There is an industry consensus that just reading news stories out loud is not good enough; an automated voice is another step backward from that.

I tried it, and I didn't like it. I just couldn't concentrate on it. I challenge you to conclude otherwise!

So yes - an inevitable innovation, but not one that does anything at all for the medium of podcasting. This technology must be ergonomic, rather than built around the convenience of the producer. It will only work for people when it has been humanised.

Tags (click tag to find related articles; click icon for feed):
industry comment |