Have you ever read my posts and thought to yourself, “this is nice, but I wish I could listen to these read by an AI voice instead”?
Well, now you can!
I have created a new podcast where you can listen to my posts. There are very few up there so far, as I am just trying this out to see how it works.
So far, I am finding it surprisingly easy, and the voice only sounds odd in a few situations.
You are probably wondering why I wouldn’t just use my own voice, and the answer is: it would be a huge amount of work.
It would take hours to make the recording and get it just right, not to mention I don’t have a studio quality microphone or a soundproof room, and am not a professional narrator.
With text to speech the voice is automatically at the exact correct volume, with no breathing or swallowing sounds to cut out, no overly loud p’s or t’s, and no background noise to worry about. Perfect studio sound quality. I only need to spend perhaps an hour or less on each episode listening to make sure it sounds correct, and making any needed changes.
One funny change I had to make was to write ‘red’ instead of ‘read’ when I wanted it pronounced in the past tense. Also I had to add many commas in unnatural places to make the voice pause appropriately. Even after that, though, I still had to edit the file in Audacity to insert pauses.
And all that was still miles easier than trying to record myself reading it.
So far there are three episodes of varying lengths: 4 minutes, 8 minutes, and 18 minutes, and using different background music and a couple different voices to see which ones work best, and what lengths/voices people prefer.
If you would like to listen, you can try an episode here, and also linked in my blogroll:
I would appreciate any feedback. I will probably try converting some of my short fiction into audio episodes as well.
We’ll see what happens, and I hope you enjoy!
Jonas David 